Primary Human Cells and Replicative Senescence

The human diploid fibroblast (HDF) was the first type of human cell used in cell senescence research (Hayflick and Moorhead, 1961), and remains the most well established human cell type for studying replicative senescence in vitro. Prior to the pioneering experiments by Hayflick and Moorhead in the early 1960s, researchers doing tissue culture believed that the limited life span of normal diploid cells in vitro was due to inadequate culture conditions. Work by Hayflick and Moorhead led to a number of key observations that argued that this limited replicative ability in vitro represented true senescence (see Dillin and Karlseder, this volume), not inadequate culture conditions. First, analysis of dozens of different fetal HDF strains revealed that the number of population doublings that these cells could reach before hitting senescence was approximately the same. Second, cryopreserved samples of the same HDF strain that were thawed out at different intervals over a large period of time all had the same replicative life span (in terms of population doublings). Third, when equal numbers of two distinguishable HDF strains, one capable of 20 more population doublings than the other, were mixed and then grown in vitro, only cells from the HDF strain capable of the additional 20 population doublings were detected at the terminal phase of culture (this observation was crucial in ruling out the possibility that infectious agents were responsible for attenuating the proliferative life span). Fourth, when cryopreserved samples of the same HDF strain were sent to be cultured at a number of different labs across the United States, all ceased proliferating at the same population doubling level.

Since the demonstration that the finite life span of HDF was due to replicative senescence, a number of other types of primary human cells have been shown to also have a finite replicative capacity in vitro, including keratinocytes (Rheinwald and Green, 1975), vascular smooth muscle cells (Bierman, 1978), endothelial cells (Mueller et al. 1980), lens epithelial cells (Tassin et al. 1979), muscle satellite cells (Decary et al. 1997), and lymphocytes (Tice et al. 1979). All of these cell types are only capable of -30-100 population doublings, depending on the cell type and age of the donor, before the onset of replicative senescence. While the growth conditions have yet to be established for the long-term culture of many types of human cells, including various types of stem cells, presently, human embryonic stem cells are the only type of normal human diploid cell that exhibits replicative immortality in vitro (Thomson et al. 1998). Since the cellular and molecular characteristics of the senescent state are best described for HDF, I will focus primarily on these cells for the rest of this section.

The two hallmark features of a senescent cell are that it is viable and has entered a state of irreversible growth arrest. Senescent HDF have been shown to be viable in vitro for periods in excess of one year (Matsumura et al. 1979). The latter hallmark feature of replicative senescence is based on a several studies which have shown that senescent HDF are refractory to a number of mitogenic and regulatory signals (Goldstein, 1990). To date, only the infection of senescent HDF with oncogenic viruses, such as the SV40 virus (Gorman and Cristofalo, 1985), has been shown to allow these cells to overcome the replicative block and resume proliferation. Primary cultures of other types of human cells, such as mammary epithelial cells, have been reported to spontaneously bypass the replicative senescence and acquire replicative immortality, but only with the accompanying loss of karyotypic stability (Romanov et al. 2001).

Senescent HDF primarily exhibit a growth arrest at the G1/S phase of the cell cycle, although a significant fraction, up to 20% for some HDF strains (Sherwood et al. 1988), have a 4N DNA content, indicating the presence of G1 tetraploid cells and/or G2 arrested cells in the senescent cell population. Other features of senescent HDF include increased frequency of karyotypic abnormalities, increased cytoplasmic volume, reduced motility, elevated RNA and protein content, increased granularity, increased frequency of lysosomal and Golgi bodies, and expression of a plethora of genes associated with the senescent state (Goldstein, 1990; Cristofalo et al. 2004). These genes include cell cycle inhibitors, such as p21 and P16INK4a which are members of the CIP-KIP family of kinase inhibitors (Noda et al. 1994; Alcorta et al. 1996); the transcription factor SAG (senescence associated gene) (Wistrom and Villeponteau, 1992); extracellular matrix (ECM)-degrading secreted metalloproteases, including collagenase, stromolysin, and tissue plasminogen activator (t-PA)(Cristofalo, 2004); promyelocytic leukemia-associated protein (PML) (Ferbeyre et al. 2000); senescence-associated beta-galactosidase (SA-P-gal) (Dimri et al. 1995); and, identified in recent studies, plasminogen activator inhibitor-1 (PAI-1), a downstream target of p53, which has been shown to be necessary to maintain cell cycle arrest in senescent HDFs (Kortlever et al. 2006). Of all these genes that are expressed in senescent HDF, SA-P-gal has proven to be the most useful biomarker of replicative senescence because of its ease of detection, both in vitro and in vivo, using a simple histo-chemical assay and because it also is a marker for replicative senescence of other types of somatic cells (Dimri et al. 1995).

Soon after the landmark experiments by Hayflick and Moorhead, a number of theories were developed to explain the finite replicative capacity of HDF. These theories could be classified into two groups: first, stochastic theories, all of which essentially purported that replicative senescence was due to the random accumulation of damage caused by extrinsic factors during replicative aging; second, genetic theories, which claimed that replicative senescence was a genetically programmed event, intrinsic to the very nature of the cell. In the 1970s, 1980s, and early 1990s, observations from a number of different studies led to increasing support for the genetic programming theories. Some of the first evidence that replicative senescence is a genetically determined process came from cell fusion experiments in the 1970s and 1980s (Norwood et al. 1974, Pereira-Smith and Smith, 1982). These elegant experiments showed that the senescent phenotype was dominant when young HDF were fused with senescent HDF (Norwood et al. 1974), or when immortal human cell lines were fused with senescent HDF (Pereira-Smith and Smith, 1982), to form heterokaryons containing both a senescent and nonsenescent nuclei. This implied that genetic factors inherited from the senescent nuclei were responsible for causing replicative senescence. Other experiments by Wright and Shay in the late 1980s and early 1990s led to a refinement of these genetic theories, when they showed that transformation of HDF with the SV40 large T antigen, allowed these cells to bypass the normal senescence checkpoint and continue to divide for an extended number of doublings (Shay and Wright, 1989). Soon afterward, however, these transformed cells underwent massive cell death (crisis). With continued culture, immortalized clonal cell populations did eventually emerge from crisis, but only at a very low frequency (1 out of every 2-3 x107 transformed cells). This suggested that were at least two genetic barriers or checkpoints in HDF that prevented replicative immortality, which they referred to as M1, the replicative senescence checkpoint, and M2, the checkpoint at crisis.

One of the most intriguing genetic program theories to explain replicative senescence was developed independently by Olovnikov (Olovnikov 1973) and Watson (Watson 1972). This theory, referred to as the end replication problem (Fig. 2.1), provided a detailed mechanistic hypothesis as to why HDF and other primary human cell cultures can only undergo a finite number of population doublings before hitting replicative senescence (see Dillin and Karlseder, this volume). This theory was based on the biochemical properties of DNA polymer-ase, the key component of the enzymatic complex responsible for replicating

Fig. 2.1 The end replication problem. The replication of the 5' strand (red) and 3' strand (blue) of a single telomere are shown, as well as adjacent nontelomeric DNA (black). Newly synthesized DNA (i.e., the daughter strand) is indicated by the dashed lines. Replication of the 5'strand occurs by lagging strand synthesis, where Okazaki fragments are shown in grey. Replication of the 3' strand occurs by leading strand synthesis. All telomeres end in a 3'overhang, which the normal DNA replication machinery cannot replicate. Degradation of the most 5' Okazaki fragment presumably allows regeneration of the overhang during replication of the 5' strand, allowing complete replication of the telomere on this daughter chromosome. However, the overhang cannot be regenerated by de novo DNA synthesis during leading strand synthesis, and therefore degradation of the 5' strand is presumably required, which is hypothesized to account for the telomere shortening observed during cell division in cells that do not express telomerase (See Color Plate)

Fig. 2.1 The end replication problem. The replication of the 5' strand (red) and 3' strand (blue) of a single telomere are shown, as well as adjacent nontelomeric DNA (black). Newly synthesized DNA (i.e., the daughter strand) is indicated by the dashed lines. Replication of the 5'strand occurs by lagging strand synthesis, where Okazaki fragments are shown in grey. Replication of the 3' strand occurs by leading strand synthesis. All telomeres end in a 3'overhang, which the normal DNA replication machinery cannot replicate. Degradation of the most 5' Okazaki fragment presumably allows regeneration of the overhang during replication of the 5' strand, allowing complete replication of the telomere on this daughter chromosome. However, the overhang cannot be regenerated by de novo DNA synthesis during leading strand synthesis, and therefore degradation of the 5' strand is presumably required, which is hypothesized to account for the telomere shortening observed during cell division in cells that do not express telomerase (See Color Plate)

DNA in dividing cells, which precluded it from completing the replication of the ends of a linear DNA molecule (i.e., chromosomes). Both of these hypotheses predicted that the ends of chromosomes would slowly erode in a dividing cell population. Olovnikov went on to further predict that the eventual deletion of essential genes located near the ends of chromosomes would ultimately cause replicative senescence (Olovnikov 1973).

0 0

Post a comment