壽Pharmacological approaches to defining the role of chaperones in aging and prostate cancer progression

The underlying mechanisms of human aging are as yet not clearly defined partly because of the complexity and the intricate pathways involved in the process. There are clearly a number of environmental and genetic challenges that contribute to a shortened cellular life span, but investigations into their interrelationships have received less attention. A diminished stress response and decreased function of the molecular chaperones hsp90 and hsp70 (Gutsmann-Conrad et al 1998) appears to contribute to replicative senescence (Lee et al 1996), which can be overcome by chronic stress and elevated chaperone levels. There is accumulating evidence to suggest that both a mild heat shock of human fibroblast cultures (Rattan 1998) and forced expression of hsp70 in invertebrates (Wheeler et al 1995) lead to enhanced tolerance to environmental stress (hormesis) resulting in increased longevity. Additional aging mechanisms that have been proposed implicate oxidative damage together with accelerated telomere erosion as causative agents of premature senescence. For example, by culturing cells in low oxygen (hypoxic conditions) or in the presence of reactive oxygen species traps, primary human cells exhibit a reduced rate of telomere shortening and an extended life span (von Zglinicki et al 2000), suggesting that culturing cells in “normoxia” (standard 20%) leads to increased telomere erosion and decreased life span. It is clear that a number of factors play important roles in the onset of senescence, from oxidative damage to a depressed stress response to telomere erosion. Consistent with these hypotheses and data suggesting that cancer is associated with an increase in chaperone function, a decline in the stress response during senescence may be critical for this proliferative block, eventually resulting in the prevention of immortalization and progression to cancer. We discuss here telomeres and telomerase as they relate to chaperones, stress response, therapy, aging, and ultimately cancer as a disease of aging.

CONNECTING TELOMERES AND THE STRESS RESPONSE DURING SENESCENCE

Normal human cells are unable to proliferate indefinitely, whether in the human body or when explanted into tissue culture. Clearly, utilization of in vivo models most accurately addresses organismal aging; yet, in vitro systems provide an appropriate model to study certain pathways related to the cellular senescence program. The proliferative checkpoint to continuous cell division has been defined as replicative senescence (Hayflick and Moorhead 1961). The senescence process appears to correspond directly with the aging of the organism as a whole, given that cells cultured from a child divide more number of times than cells from an aged adult. In addition, as human cells approach senescence, they show a severe loss of responsiveness to environmental stress. Heat shock followed by recovery has been shown to provide a sustained stress response and an extension of cellular life span (Rattan 1998), presumably via a cytoprotective mechanism known as hormesis. Hormesis is defined as the adaptation or tolerance of cells or organisms to repeated mild stress in order to survive certain types of lethal environmental challenges. Additional data suggest that chaperone introduction into invertebrate animals results in an extension of life span during organismal aging (Wheeler et al 1995). Recent results in Drosophila indicate a protective role for chaperone function in the prevention of neurodegenerative disorders, such as Parkinson's disease (Auluck et al 2002), further suggesting the decreased function of chaperones during aging.

Our current understanding directly relates the onset of senescence with the shortening of telomere repeats, specialized structures at the ends of the chromosomes. Each time a cell divides telomeric deoxyribonucleic acid (DNA) is lost because of the inability of conventional DNA polymerases to replicate to the end of the linear chromosome (Harley et al 1990). Amazingly, the introduction of the enzyme responsible for telomere maintenance, telomerase, blocks telomere erosion in normal cells, ultimately preventing cellular senescence (Bodnar et al 1998; Vaziri and Benchimol 1998) without signs of cancer-associated changes (Jiang et al 1999; Morales et al 1999). Thus, although there are a variety of theories and hypotheses on human aging, including increased oxidative damage or depressed response to environmental stress (or both), these data provided the first direct experimental evidence that telomere shortening is a primary cause of cellular senescence. Consistent with these observations are studies in telomerase knockout mice that show telomere shortening is directly related to organismal aging (Rudoph et al 1999), indicating that in vitro human models are an appropriate substitute for defining mechanisms related to aging in humans. It seems reasonable to suggest that human cells exposed to a sustained stress response or increased expression of chaperones would exhibit slower rates of telomere shortening compared with unstressed cells. However, maintaining a stress response may instead reduce the threshold level where telomeres are recognized as damaged DNA. Alternatively, because it has been recently shown that the shortest telomere is responsible for senescence and not the overall telomere lengths (Hemann et al 2001), it is possible that in cells with a sustained stress response, shortening of the shortest telomere may not occur at the same rate as untreated cells, leading to life span extension.

The interplay between stress, telomeres, telomerase, aging, and diseases of aging (namely, cancer) has yet to be experimentally clarified. The telomerase enzyme is minimally composed of a protein component, human telomerase reverse transcriptase (hTERT), the catalytic subunit (Meyerson et al 1997; Nakamura et al 1997), and a template ribonucleic acid (RNA), human telomerase RNA component (hTR), which is responsible for recognizing the 3′ overhang of the chromosome ends to template telomere addition (Feng et al 1995). The fact that telomerase migrates as a large molecular weight complex in glycerol gradients suggests that a number of proteins associate with the core components of telomerase (Bednenko et al 1997). Although there are a number of identified proteins that associate with telomerase, the chaperone proteins hsp90, hsp70, and p23 were the first shown to functionally associate with telomerase and to be required for telomerase assembly in vivo and in vitro (Holt et al 1999). Proteins hsp90 and p23 appear to remain associated with active telomerase to aid conformational changes that may take place as the enzyme translocates along the telomere to catalyze the addition of the next telomeric repeat. Data indicate that hsp70 interacts with hTERT, but unlike the stable association of hsp90 and p23, it dissociates upon proper conformation and activation of telomerase (Forsythe et al 2001). The minimal components required to assemble the human telomerase were identified as hTERT, hTR, hsp90, p23, hsp70, hsp40, and the heat shock organizing protein, HOP/p60 (Holt et al 1999). To date, the roles of the hsp40 and HOP proteins in telomerase assembly and regulation have not been determined but may be related to the kinetics of the assembly process. Given that telomere erosion, telomerase activity, and chaperones appear functionally interrelated, it is reasonable to assume that there may be a link between DNA damage, stress response, and telomere shortening related to the onset of replicative senescence or the immortalization-transformation process (or both).

PROSTATE CANCER THERAPY: THE UNIQUE TARGETING OF CHAPERONE FUNCTION

As a condition of the aging process, cancer may have a unique place as a disease where chaperone proteins play an integral role in therapeutic intervention. We have begun to dissect out the mechanisms involved in prostate cancer progression using an appropriate model system within a similar genetic background. Nonneoplastic prostate epithelial cells obtained from a radical prostatectomy were immortalized using SV40 large T antigen and shown to be nontumorigenic in the athymic nude mouse under standard conditions. However, after a latent 6-month incubation of the immortalized P69SV40T cells in the athymic mouse, 2 out of 18 mice developed sporadic tumors during an in vivo selection process, followed by in vitro culture and subsequent subcutaneous reinjection into mice to produce metastatic sublines. Therefore, the model system consists of cell lines from nontumorigenic (P69), tumorigenic (M2182), and metastatic (M12) cells of the same genetic lineage and cytogenetic aberrations, consistent with authentic human prostate cancer (Bae et al 1994, 1998).

Our studies with this model indicated that there is a dramatic increase in telomerase (∼15-fold) from the nontumorigenic cells (P69) to the tumorigenic (M2182) and metastatic (M12) cells, again supporting the critical role of telomerase in cancer progression (Akalin et al 2001). We also detect a substantial increase in chaperones and chaperone-related proteins in tumorigenic and metastatic cells, which directly correlates with the observed elevation in telomerase levels. These increases were found to be directly related: the change in telomerase activity levels was almost exclusively caused by increased chaperone-mediated assembly rather than elevated expression of the telomerase components. This is the first known model system to suggest the importance of chaperone proteins and telomerase in the transformation process in prostate cancer and identifies a novel, potentially important pathway for chaperone-mediated telomerase assembly during cancer progression. To determine if the increase in chaperone proteins is biologically significant, we have immunohistochemically analyzed nearly 75 primary prostate biopsy cases and found that chaperone up-regulation is consistently associated with tumorigenic conversion to prostate carcinoma (in preparation). Our data indicate that the levels of hsp90 and p23 expression are consistently (greater than 95% of the cases) higher in prostatic intraepithelial neoplasia and prostate carcinoma, relative to benign prostatic hyperplasia (in preparation). Because we find a dramatic increase in chaperones and chaperone-related proteins during tumorigenic conversion both in vitro and in vivo, these proteins have become novel targets for chemotherapeutic prostate cancer intervention.

Accumulating evidence suggests that certain hsp90 inhibitory compounds block the chaperone-mediated assembly of functional enzyme complexes. Of these drugs, radicicol and geldanamycin and its derivatives have shown the most promise. Radicicol is an antifungal antibiotic that was shown to inhibit transformation by oncogenes v-src, ras, raf, and mos (Kwon et al 1992; Zhao et al 1995). Isolated from the fungus Monosporium bonorden, radicicol is the prototype of a second class of hsp90 inhibitors that is unrelated structurally to the first class of inhibitors, the benzoquinone antibiotics (eg, geldanamycin and herbimycin A) (Schulte et al 1998). All of these hsp90 inhibitors have been shown to bind to the N-terminal adenosine triphosphate (ATP)–adenosine diphosphate (ADP) binding site of hsp90, where they compete for ATP binding and trap the molecule in the ADP-bound conformation, thereby inhibiting hsp90 function (Grenert et al 1997; Roe et al 1999; Schulte et al 1999). Radicicol has been shown to bind to all the hsp90 family members, including hsp90 α and β, Grp94 (the hsp90 homolog in the endoplasmic reticulum), and the newly discovered Trap-1 (hsp75), although it has a greater affinity for hsp90 (Schulte et al 1999). Recent studies suggest a limited clinical usefulness of these drugs because of their hepatotoxicity and narrow therapeutic index (Supko et al 1995); yet, newer synthetic analogs, like the geldanamycin analog 17-(allylamino)-17-demethoxygeldanamycin (17AAG), are proving to have a better therapeutic index (Burger et al 1998). In addition, other drugs (novobiocin and related coumarins) that are reported to target hsp90 are now being identified (Marcu et al 2000a, 2000b) and are already established as well-tolerated compounds.

Our initial studies using the prostate cancer progression model suggest that we can successfully block telomerase assembly using chronic treatments of the hsp90 inhibitor radicicol to target telomerase activity therapeutically. At subtoxic concentrations that do not significantly affect cell growth rates, we are able to dramatically reduce the telomerase activity (Fig 1), which is an indicator of chaperone inhibition. Our aim is to further extend these encouraging results in terms of telomere shortening and induction of senescence or cell death and to identify additional compounds that will serve as prostate cancer inhibitors. We believe that we can target the excess or elevated functional chaperones observed in the tumorigenic cells to deplete telomerase activity, triggering telomere shortening and reprogramming of replicative senescence or induction of cell death. This novel treatment strategy may prove invaluable in the treatment of cancers, including prostate and breast. Although such a novel therapy for prostate cancer is warranted, it would likely be an adjuvant treatment, with general tumor resection as the first line of defense. Treatments for chaperone inhibition as a means to block telomerase assembly would be useful to aid in preventing the repopulation of the more aggressive cancerous cells left behind after the primary treatment (ie, tumor resection). It is widely believed that the primary tumor in many cases is not fatal; yet, the recurrence of a secondary tumor with invasive and metastatic capabilities results in increased mortality rates. On the basis of our recent data, specific inhibition of telomerase or telomerase-associated proteins as an adjuvant therapy for preventing recurrence of disease, especially in prostate cancer, may be a realistic objective.

Fig. 1.

 Specific inhibition of the chaperone-mediated telomerase assembly with chronic radicicol treatment of the prostate cancer cells. The tumorigenic, metastatic prostate cancer cell line, M12, was treated with 0.3 μM and 5.0 μM radicicol (RAD) for the indicated days and tested for the inhibition of telomerase activity using the Telomeric Repeat Amplification Protocol (TRAP) assay (Kim et al 1994). A characteristic 6-bp telomerase-specific laddering is observed in the presence of telomerase activity (eg, lane 1), whereas the lysis buffer only (Buffer) serves as an appropriate negative control. The 36-bp internal standard serves as a useful internal control for the normalization of sample to sample variation, as well as for semiquantitative purposes. Quantitation (Q) is accomplished by obtaining the ratio of the telomerase ladder to the internal control, with untreated samples normalized to 100%. Although growth rates of the M12 cells were not affected by 0.3 μM radicicol, a gradual and specific inhibition of the telomerase activity was observed after long-term treatment. A 5.0-μM treatment was cytotoxic and caused substantial cell death (not shown) with an obvious decline in telomerase activity

Interestingly, recent studies have suggested that geldanamycin-mediated inhibition can be directed at specific targets. For example, geldanamycin-testosterone hybrids have been shown to have selective activity toward cancer cells that express the androgen receptor (Kuduk et al 2000), whereas geldanamycin dimers (ie, GMD-4c) target HER-kinase in HER-kinase overexpressing cancer cells (Zheng et al 2000). These synthetic molecules may prove to be useful in identifying novel approaches for the inhibition of specific hsp90-protein interactions without the toxic side effects associated with targeting all hsp90 targets in all cell types. As more compounds are identified as hsp90 inhibitors, the discovery of more specific and less toxic chemotherapeutic drugs that are well tolerated in humans will likely follow.