Variation in cytoskeletal assembly during spreading of progressively subcultivated human embryo fibroblasts (IMR-90)

The concept of telomeric non-reciprocal recombination (TENOR) applied to human fibroblasts grown in serial cultures: concordance with genealogical data

Since the discovery of the limited life span of human fibroblasts some 50 years ago, many genealogical studies have been undertaken to describe growth kinetics of fibroblasts in serial cultures by their individual division behavior. It is now accepted that proliferation capacities of human fibroblasts strongly depend on their telomere lengths and integrity. Telomeres shorten with each replication round, and there is a direct correlation between cell division capacity and telomere lengths; that is, the consumption of disposable telomeric DNA repeats during cell divisions progresses until critically short telomeres determining the replicative senescence of the cells are present. Recently, we have suggested that telomeres in fibroblasts can also become elongated during DNA replication by telomeric non-reciprocal recombination (TENOR). Here we discuss genealogical data collected over the last decades as well as more recent findings on the telomere-driven replicative senescence process, and we summarize both to give an integrated picture.

Changes in three-dimensional architecture of microfilaments in cultured vascular smooth muscle cells during phenotypic modulation

To investigate changes in the three-dimensional microfilament architecture of vascular smooth muscle cells (SMC) during the process of phenotypic modulation, rabbit aortic SMCs cultured under different conditions and at different time points were either labelled with fluorescein-conjugated probes to cytoskeletal and contractile proteins for observation by confocal laser scanning microscopy, or extracted with Triton X-100 for scanning electron microscopy. Densely seeded SMCs in primary culture, which maintain a contractile phenotype, display prominent linear myofilament bundles (stress fibres) that are present throughout the cytoplasm with alpha-actin filaments predominant in the central part and beta-actin filaments in the periphery of the cell. Intermediate filaments form a meshed network interconnecting the stress fibres and linking directly to the nucleus. Moderately and sparsely seeded SMCs, which modulate toward the synthetic phenotype during the first 5 days of culture, undergo a gradual redistribution of intermediate filaments from the perinuclear region toward the peripheral cytoplasm and a partial disassembly of stress fibres in the central part of the upper cortex of the cytoplasm, with an obvious decrease in alpha-actin and myosin staining. These changes are reversed in moderately seeded SMCs by day 8 of culture when they have reached confluence. The results reveal two changes in microfilament architecture in SMCs as they undergo a change in phenotype: the redistribution of intermediate filaments probably due to an increase in synthetic organelles in the perinuclear area, and the partial disassembly of stress fibres which may reflect a degradation of contractile components.

Chromatin reorganization during senescence of proliferating cells

It was previously proposed (Macieira-Coelho, 1979) that aging of proliferating cells is the result of genome reorganization taking place during the division cycle. This hypothesis was investigated and a reorganization could indeed be ascertained in the different hierarchical orders of DNA structure; a correlation was found between changes in chromatin organization and the impairment of cell cycle-related events. Indeed, like the latter, the reorganization of chromatin structure is characterized by a succession of subtle changes through the cell population life span, and a final short stage with abrupt events. The final events seem to concern mainly the organization of heterochromatin. The reorganization in the genome is accompanied by structural changes in the cellular scaffold and an evolution of cell morphology. The remodeling occurring in the cell through serial divisions seems to take place in such a way as to decrease the probability of further reorganizations, tending to a limit. The decline of the proliferative activity seems to be the result of the tendency to reach this limit.

Changes in the cell surface of human diploid fibroblasts during cellular aging

The electrophoretic mobility of 13 human diploid cell strains, TIG-1, TIG-2, TIG-3, TIG-7, WI-38, IMR-90, MRC-5, MRC-9, TIG-1H, TIG-1L, TIG-2M, TIG-2B, and TIG-3S, which were established from different tissues of human embryos, was studied at different passages. The net negative surface charge of the cells was characteristic for each cell strain and decreased significantly during the in vitro aging of the cells. The decrease in the net negative charge of the cells correlated well with the decrease in cell density throughout the life span of the cells. A strict linear correlation between the electrophoretic mobility and the number of cells harvested at each passage was obtained for all the human diploid cell strains. Moreover, almost the same linear regression coefficient of the cells was obtained among these cell strains. Therefore, the net negative surface charge of human diploid cell strains could serve as a cell surface marker for in vitro cellular aging.

Correlation between contractility and proliferation in human fibroblasts

The contractile power of human fibroblasts was checked through their life span in vitro, using a plasma clot retraction test. It was found to decline with a pattern analogous to that of the different phases identifiable by the study of the kinetics of proliferation of these cells. The capacity to retract a plasma clot was higher in cells harvested during active growth than in cells harvested in resting phase. The decreased ability to retract during aging becomes apparent when cells are harvested in resting phase. Decreased retractile activity was also observed in postnatal cells as compared with embryonic cells. The results support a correlation between the initiation of DNA synthesis and the turnover of cytoskeletal elements. The data fit our previous results showing that the early proliferative disturbance during cellular senescence consists of a decline in the probability of initiating the division cycle linked to impaired cell attachment and spreading.

On the relationships between the rate of cytoskeletal stable assemblies turnover, stability of the differentiated state, development and aging

There is a general consensus that biological specificity is a structure-derived property. If a living system is going to maintain its structure and function then the newly synthesized molecules should replace the faulty ones at the correct time and in the correct places so that the previously established cellular topology will be preserved. In addition, pre-existing spatial determinants which will direct the asymmetrical assembly of the newly synthesized molecules should be available. Therefore, regulation of turnover of cellular architecture represents an essential feature of living systems. In considering the underlying causes of cellular senescence it seemed reasonable to focus on the relationship between development of a stable phenotype and the turnover of cellular and extracellular stable assemblies, currently thought to be involved in maintaining the stability of the differentiated state. In recent years evidence has accumulated suggesting a reciprocal relationship between cytoarchitecture turnover rate and achievement of a stable structure. The lack of a feedback control on the turnover of cellular stable assemblies and/or a low turnover rate of cytoarchitecture components would mean that they will be subjected to damaging processes such as oxidation, cross-linking, aminoacid racemization or non-enzymatic browning which are known to occur in other long-lived proteins. The consequence would be the generation, with advancing age, of faulty cellular structures which, in turn, would alter the deposition of newly synthesized molecules. This process may lead to a progressive breakdown in cellular and extracellular stable structures. The process of directed assembly seems to be general for biological systems displaying history-dependent development. We believe that it is this strategy which imposes severe limitations on presegregated spatial determinants turnover rates and, therefore plays a major role in initiating the aging process. We also suggest that species-specific life-span might be determined by the species-specific regulatory networks which governs the cell-specific cytoarchitecture damaging rate. Moreover, aging appears to be an intrinsic feature of biological systems displaying history-dependent development and should be absent in systems displaying history-independent life-cycles, such as bacteria, some species of protozoa, and certain transformed cell lines. An important feature of protein turnover is that this process requires metabolic energy. Therefore, we can expect that structure preservation strategy is a part of a more general energy-saving strategy, a view previously expressed by T.B.L. Kirkwood (Nature, Lond., 1977, 270, 301-304).

Effect of donor age and prior sun exposure on growth inhibition of cultured human dermal fibroblasts by all trans-retinoic acid

The effect of retinoic acid on human fibroblasts was studied in a cell culture model of chronologic aging and photoaging. During early exponential phase, all trans-retinoic acid significantly stimulated growth rate of adult arm-derived dermal fibroblasts but not of newborn or adult foreskin-derived fibroblasts. Retinoic acid also significantly reduced saturation density in most young adult arm-derived lines and all 24 lines derived from old adult arm and foreskin. However, four of ten young adult outer arm (relatively sun-exposed) and one of ten young adult inner arm (relatively sun-protected) fibroblasts lines increased their saturation density in response to retinoic acid. These data suggest that prior sun exposure and/or donor age may influence cellular responsiveness to retinoic acid. Neither the stimulatory nor the inhibitory effect of retinoic acid could be attributed to cell density, to breakdown of retinoic acid in culture, to nutrient depletion or to serum dependency. However, stationary phase fibroblasts from all sites (foreskin, inner and outer arm) showed an increase in filopodia and in intracellular actin after treatment with retinoic acid that was roughly proportional to the degree of growth inhibition, irrespective of donor age. We suggest that retinoic acid induces premature density dependent growth inhibition at least in part by increasing filopodia-mediated cell contact that is in turn directly related to an increase in fibrillar actin.

Gene reorganization during serial divisions of normal human cells

We have followed during serial divisions of human fibroblasts the presence in chromosomal and extrachromosomal DNA, of two genes that are expressed in fibroblasts, actin and interferon, and of one that is not expressed, globin. The intensity of the blot hybridization of the actin and globin probes with chromosomal DNA diminished during serial divisions of diploid fibroblasts. The interferon gene remained constant throughout the human fibroblast life span. Chromosomal DNA sequences were present in extrachromosomal circular DNA which appeared at the end of the fibroblast life span. The results could explain some functional changes that occur in these cell populations when their division potential declines.

Fibronectin localization and endocytosis in early and late mouse embryonic fibroblasts in primary culture: a study by light and electron microscopic immunocytochemistry

In spreading fibroblasts, strong endocytosis of exogenous fluorescent fibronectin (FN) was observed from the beginning of their attachment to the substratum. In early fibroblasts, the internalized FN was localized both in the peripheral ruffles and in the perinuclear cytoplasm; in late fibroblasts, whose spreading was slower, FN uptake was not detected in the ruffles. In growing cultures, supracellular FN fibres, detected by direct fluorescence microscopy or by the indirect peroxidase-anti-peroxidase (PAP) complex technique, were scarce on early cells, but very numerous on the upper face and on the filopodia of late cells. At the ultrastructural level, FN, localized with the immuno-gold staining method, was found associated with fibres of the extracellular matrix and the upper face of the cells. FN was endocytosed via smooth vesicles and we suggest that the internalization process is slower in the late cells. In confluent early cultures, an extended network of pericellular FN was observed as usual. The pericellular FN of late grouped cells was present as a few coarse fibres connected with some of the cell surface threads.

Reduction of filamin in late passage human diploid fibroblasts (IMR-90)

Progressive subcultivation of IMR-90 cells results in non-proliferative, heterogeneous cultures which may reflect aging of the diploid line (Hayflick, Exp. Cell Res., 37 (1965) 614). We have observed that late passage cells exhibit different rates of spreading and morphogenesis when compared to early passage groups, phenomena which we attribute to altered reassembly of the cytoskeleton in senescent cells (Kelley et al. Mech. Ageing Dev., 13 (1980) 127). To determine whether potential differences in cytoskeletal proteins develop with progressive subcultivation, early and late passage cultures were extracted with 0.5% Triton X-100 for 1 min followed by 1.0% sodium dodecyl sulfate (SDS) prior to separation and characterization of extracted proteins by electrophoresis on 7.5-15% gradient SDS gels. Extractions were made of both culture groups 3, 6 and 24 h after reseeding. Cytoskeletal ultrastructure at each stage of spreading was examined either in replicas of extracted cells or directly by scanning electron microscopy. Although considerable variation in cytoskeletal organization was observed, qualitative differences in gel banding patterns of actin, myosin and tubulin were not apparent at selected time points. However, late passage cells at 6 h and 24 h did not exhibit filamin associated with the Triton insoluble fraction as did early passage cells. Since it has been demonstrated that filamin is capable of cross-linking actin microfilaments into bundles or sheets, we suggest that it is a principal element for the variant cell shape and cytoskeletal morphology observed during altered spreading behavior of late passage human diploid fibroblasts.

A simple technique for the visualization of whole mount cytoskeletons with transmission electron microscopy

Examination of whole mount cells in the transmission electron microscope has been useful in studies of cellular architecture. The common technique is to grow cells directly on formvar-coated, gold grids for direct observation through a cell. We report a technique for obtaining whole mount preparations which requires neither fragile formvar films nor expensive, gold grids. Cells are grown on palladium-coated coverslips and processed for electron microscopy. The cells and the palladium substrate are separated from the coverslip. The cell-palladium complex is then picked up on copper grids as in thin section processing. We compare images of the cytoskeleton using our technique with images using previously described techniques and present preliminary observations of contracting cell models. Such contractions would tear formvar films if attempted on cells grown in the conventional manner for whole mount examination. Our technique allows cells to contract without tearing the underlying substrate.

Isolation by flow sorting of cytokinetic and morphological heterogeneity in late-passage cultures of human diploid fibroblasts (IMR-90)

Considerable structural, metabolic, and proliferative heterogeneity develops in populations of cultured diploid cells which have reached advanced levels of population doubling. Isolation of noncycling cells from late-passage cultures would permit more definitive investigation of the structure and behavior of individual senescent cells. In this paper, we report the viable sorting of late-passage cultures of human diploid fibroblasts (IMR-90) into two subpopulations of cells with different proliferative potentials. Sorting is based on cellular light-scattering properties and autofluorescence. Structural and behavioral features of the subpopulation exhibiting increased forward-angle light scatter are more characteristic of senescent cells than the subpopulation sorted by decreased forward-angle light scatter.

Organization of the cytoskeleton in square fibroblasts

The relationship between the cytoskeleton, stress fiber formation, and cell shape has been difficult to determine in fibroblasts grown in tissue culture. Vagaries in cell shape are complicated, as well, by stochastic cell movements. We dictated the attachment sites and shape of fibroblasts by growing them on square adhesive substrates surrounded by nonadhesive substrates. Cytoskeletal models were made by treating the cells with buffered Triton X-100 and glycerol. The residues were then examined by scanning electron microscopy followed by light microscopy of the same cells. The cytoskeletons of randomly moving cells were examined with whole mount transmission microscopy to confirm images seen with scanning microscopy. The cells thus examined demonstrated definite relationships between ruffling activity and stress fiber terminations, which were limited to the more adhesive, palladium substrate. No stress fibers were seen to end on the lesser adhesive substrate, agarose, and ruffling did not occur across the agarose. Cells too small to fill an entire square tended to extend across one diagonal of the square, and the stress fibers ran parallel to the longest axis of these cells. Larger cells were able to completely fill their squares. The cytoskeletons of these cells were organized in a spatial relation to the square shape of the cells. The cortical meshwork was aligned circularly and diagonally within the cells. Stress fibers appeared to form from the microfilaments of the meshwork and were aligned diagonally across the cells. We conclude that the diagonal arrangement of the stress fibers and cortical meshwork is caused by the same mechanism by which smaller cells spread over the longest axis of a square. Regions of cells where the meshwork was absent or where stress fibers were tightly bundled were occupied by more randomly arranged cytoskeletal components. Regions of tightly bundled stress fibers did not seem to coincide with regions of cortical meshwork as seen by either whole mount transmission or scanning electron microscopy. Stress fibers were revealed in the light microscope to course beneath more randomly oriented cytoskeletal elements. These "lacework-like" elements were found frequently in square cells. Conspicuous structures in this random lacework were focal points of radially arranged filaments. Our observations suggest a continuity between stress fibers and the cortical microfilaments. The orientation of fibers and filaments was, in turn, dependent on cell shape for organization within the cell.

Changes in nuclear shape and mitochondrial structure do not accompany the loss of division potential in human fibroblasts in vitro

Previous studies on ultrastructural changes that occur in cultured human fibroblasts during their in vitro life-span indicate that "senescent" cells characteristically possess structurally altered mitochondria, highly lobed nuclei, and an abundance of secondary lysosomes when compared to early passage cells. In the present study, we demonstrate that improper preparative methods can induce altered mitochondrial morphology in preparations of both IMR-90 and HF730A fibroblasts, regardless of passage level. We also show that nuclei of both living and fixed IMR-90 fibroblasts are ovoid in shape, not lobulate, in well-spread cells, regardless of either the passage level or the proliferative capacity of the cell. Fibroblasts contain lobulated nuclei only when they have not spread completely on the culture substrate. Lobulations can be induced at any passage level by collagenase/trypsin or trypsin/EDTA treatment prior to fixation, but not by cytochalasin B treatment or by cold temperatures. We conclude that any treatment that affects cytoskeleton-membrane-culture substrate interactions will induce this aberrant nuclear morphology, but that this is not indicative of "senescence" and does not relate to proliferative decline.

Loss of organized fibronectin matrix from the surface of aging diploid fibroblasts (IMR-90)

Indirect immunofluorescence was used to investigate the production of a fibrillar fibronectin matrix by human diploid fibroblasts (IMR-90) as cells progress through their in vitro lifespan. Early and mid-passage cultures displayed a prominent fibrillar reticulum over the cell layer which formed within 24 hours of seeding. Even sparsely seeded early-passage cells exhibited fibrils of fibronectin on external surfaces. In contrast, fibrillar fibronectin was reduced or absent on surfaces of late-passage cells. However, the larger, non-proliferating, late-passage cells were producing fibronectin, 1h determined by radioimmunoassay of the medium.

Glycosaminoglycan synthesis and composition in human fibroblasts during in vitro cellular aging (IMR-90)

The synthesis and turnover of sulfate-labeled glycosaminoglycans(35S-GAGs) has been investigated in diploid human embryo fibroblasts during in vitro cellular aging. With progressive subcultivation, there was a decreased incorporation of Na2(35)SO4 into 35S-GAGs released to the medium, but not into those accumulated at the cell surface. The composition of 35S-GAGs found in extracellular medium, cell surface (removable by gentle proteolysis), and intracellular compartments of the culture after 48-hr labeling did not change significantly with progressive subcultivation. Pulse-labeled 35S-GAGs moved from intracellular to surface and extracellular compartments more slowly in late-passage cultures. Addition of 1 mM beta-xyloside to both early- and late-passage cultures produced a ten-fold enhancement of extracellular 35S-GAG production without a concomitant increase in surface-associated 35S-GAG. We interpret the data of this study to mean that secreted and cell-surface glycosaminoglycans represent different pools and that cellular aging has its effect primarily upon the secreted pool of glycosaminoglycans. Late-passage fibroblasts demonstrate marked decreases in proliferation, culture density, fibronectin matrix, and gap-junction formation. Our results suggest that glycosaminoglycan synthesis and composition are not intimately related to these parameters.