Cancer Therapy Vol 1, 173-178, 2003.
Can mortalin be a candidate
target for cancer therapy?
Renu Wadhwa,
Kazunari Taira and Sunil C Kaul
National Institute of Advanced Industrial Science
& Technology (AIST), 1-1-1 Higashi, Tsukuba, Ibaraki 305-8562, Japan
__________________________________________________________________________________
*Correspondence: Renu Wadhwa, Gene
Function Research Center, National
Institute of Advanced Industrial Science and Technology (AIST), 1-1-1 Higashi,
Tsukuba, Ibaraki 305-8562, Japan; Phone: +81 29 861 9464; Fax: +81 29 861 3019,
e-mail: renu-wadhwa@aist.go.jp
Key Words: mortalin,
localization, p53, inactivation, tumors, therapy
Summary
Differential staining pattern of
mortalin (mot-2/mthsp70/PBP74/GRP75) is a reliable marker for normal and
cancerous phenotype of cells. It is an essential protein, sojourns multiple
subcellular sites while residing predominantly in mitochondria. It has multiple
binding partners and performs multiple functions including mitochondrial
import, intracellular trafficking, receptor internalization and inactivation of
the tumor suppressor protein p53. The present article updates our understanding
on its functions in cellular senescence and immortalization and proposes its
use as a target for cancer therapy.
Mortalin is a member of hsp70 family of proteins.
It was first cloned from the cytoplasmic fraction of normal mouse fibroblasts;
the immortal cells lack this protein in their cytoplasmic fraction (Wadhwa et al, 1993a). Subsequently, by
immunostaining its differential subcellular distribution was recognized in
normal and immortal mouse cells (Wadhwa et al, 1993b). A large variety of human
normal and immortal cells were demonstrated to have pancytoplasmic and
perinuclear cellular distribution of mortalin, respectively. As discussed
below, its mutliple subcellular sites and binding partners signify its multiple
roles, some of which are crucial for continued proliferation of cells.
Immunostaining
of normal and immortal cells with a mortalin specific antibody revealed that it
is widely distributed in the cytoplasm of normal cells and is restricted to the
perinuclear region in immortal mouse cells. cDNAs encoding the cytoplasmically
distributed protein (mot-1) and the perinuclear protein (mot-2) were cloned
from normal and immortal mouse fibroblasts, respectively (Wadhwa et al, 1993c). These were shown to be different by two amino acids (Wadhwa et al, 1993c), have contrasting biological activity and coded by two
alleles in mouse (Kaul et al, 2000b). The mot-1 cDNA encoding the pancytoplasmic form of mouse
mortalin when introduced into NIH 3T3 cells induced cellular senescence like
phenotype in these cells (Wadhwa et al, 1993c). In contrast, the mot-2 cDNA that encoded perinuclear
protein resulted in malignant transformation of NIH 3T3 cells (Kaul et al, 1998).
Differential
cellular distribution of mortalin in normal and transformed cells was also
endorsed by human system. In more than 50 different human immortal cell lines
examined, mortalin was observed as nonpancytoplasmically distributed (Wadhwa et al, 1995) (Figure 1) in contrast to the normal cells that showed
pancytoplasmic staining. Cloning and analyses of mortalin cDNA from normal and
transformed human cells, however, revealed no significant difference (Kaul et al, 1998) proposing that there are, at least, two mechanisms
operating for differential distribution of mortalin in normal and transformed cells.
One is by distinct cDNAs, mot-1 and mot-2, and is found in mouse. The other may
involve protein modifications, binding partners or other cellular factors and
operates in mouse and human. Such mechanism(s) remains to be elucidated. Human mortalin cDNA clone when
expressed in mouse immortal cells led to their malignant transformation similar
to the one caused by mouse mot-2 cDNA. Both mouse mot-2 cDNA and human mortalin
also led to lifespan extension of normal human fibroblasts (Kaul et al, 2000a). Based on these functional data, human mortalin cDNA was
called hmot-2 and its overexpression was suggested to have pro-proliferative
function.

Figure 1. Mortalin immunostaining in normal (skin fibroblasts, TIG-1) and transformed (osteogenic sarcoma, U2OS) human cells.
Subcellular
distribution of mortalin shifted from the perinuclear to the pancytoplasmic
type when cancerous cells were induced to senescence. For example, introduction
of human chromosome 7 to carcinogen-transformed liver fibroblasts (SUSM-1)
resulted in their senescence in culture as determined by their proliferation,
senescence associated b-gal activity. The senescent cells showed
pancytoplasmic distribution of mortalin (Nakabayashi et al, 1999). In a similar approach when
chromosome-fragments and genes from human chromosome 4 were introduced into
cervical carcinoma (HeLa) cells, resulted in an induction of senescence
phenotype that was accompanied by a shift in the subcellular distribution of
mortalin from a perinuclear to pancytoplasmic type (Bertram et al, 1999). Induction of senescence like growth
arrest by bromodeoxyuridine (Michishita et al, 1999) or MKT-077 (a rhodacyanine dye that is
selectively toxic to cancer cells) also caused shift of subcellular
distribution of mortalin from perinuclear to pancytoplasmic type,
characteristic of normal cells (Wadhwa et al, 2000). On the other hand, Simian Virus 40 large
T antigen (SV40 LTAg) - induced cellular transformation of human lung
fibroblast (MRC-5) cells resulted in shift of pancytoplasmic mortalin staining
in normal cells to the nonpancytoplasmic staining in its immortal derivatives.
Taken together, these studies showing the absence of pancytoplasmic mortalin
staining pattern in cancerous cells have assigned mortalin staining as a
reliable marker of cellular normal and transformed phenotypes.
Mortalin is a highly conserved member of hsp70 family
of proteins. It was also cloned as a peptide binding protein (PBP74) (Dahlseid et al, 1994; Domanico et al, 1993), mitochondrial heat shock protein 70 (mthsp70)
(Bhattacharyya et al, 1995) glucose regulated protein 75 (GRP75) (Webster et al, 1994) and was found in multiple subcellular sites by a
variety of protocols. Confocal laser microscopy of the native protein with
protein-specific antibodies in a variety of cell lines revealed its existence
in multiple subcellular sites that includes mitochondria, endoplasmic
reticulum, cytoplasmic vesicles and cytosol (Dahlseid et al, 1994; Domanico et al, 1993;
Poindexter et al, 2002; Ran et al, 2000; Singh et al, 1997; Soltys and Gupta,
1999, 2000; Wadhwa et al, 1995; Webster et al, 1994). These data suggest its involvement in multiple
cellular functions.
In support to its localization at multiple subcelluar
sites, mortalin was shown to bind to residents of different organelles by a
variety of protocols. Far-western screening identified glucose regulated ER
chaperone (GRP94) as one of its binding partners. Mortalin-GRP94 interactions were
confirmed by mammalian two-hybrid assays, in vitro and in vivo coimmunoprecipitations (Takano et al, 2001). Mortalin
was also isolated as FGF-1 binding protein by FGF-1 affinity chromatography and
was shown to aid in its intracellular trafficking (Mizukoshi et al, 1999), mediated by its cell cycle specific phosphorylation (Mizukoshi et al, 2001). ATP-sensitive association of mortalin with IL-1
receptor type was also detected and predicted to have a role in receptor
internalization (Sacht et al, 1999). Yeast interactive
screen for mortalin binding proteins isolated the mitochondrial proteins hsp60,
NADH dehydrogenase, Tim44, Tim23 (unpublished data) and the peroxisomal protein
MPD (Wadhwa et al, 2003a) as its binding partners. It appears that mortalin
routes through multiple subcellular sites and thus interacts with different
proteins therein. Recently, it has been recognized that protein distribution in
a cell is more dynamic than was earlier thought. Many other proteins have been
detected in subcellular localizations that were considered foreign previously (Soltys and Gupta, 1996, 1997, 1999). The studies warrant further analyses to elucidate
the kinetics of mortalin binding to its binding partners, their temporal and
special relevance to cellular mortal, immortal and stressed phenotypes
including apoptosis.
Mortalin
binds to p53 in transformed human cells (Wadhwa et al, 1998). These interactions result in cytoplasmic sequestration of
p53 and inhibition of its transcriptional inactivation function (Wadhwa et al, 1998; 2002b). Such inactivation of p53 function could account for
lifespan extension of normal human cells. Recently, it has been shown that
telomerase in cooperation with mortalin could accelerate the immortalization of
normal human cells (Kaul et al, 2003). Binding studies using deletion mutants have demonstrated
that an N-terminal region of mortalin binds to the C-terminus of p53,
previously shown to be involved in cytoplasmic sequestration of p53 (Kaul et al, 2001; Moll et al, 1992; Wadhwa et al, 2002b). Most recently, it has been shown that p53 also exists in
the mitochondria (also a predominant localization of mortalin) and interacts
with mortalin/mthsp70, Bcl-2 and hsp60 (Dumont et al, 2003; Mihara et al, 2003) and these
interactions are involved in p53-mediated apoptosis by a pathway
independent to its nuclear function. In this scenario, if mortalin could
interfere with the p53-Bcl-2 interactions it may act as an antiapoptotic factor
(Figure 2). Such possibilities remain
to be tested. On the other hand, abrogation of moratlin-p53 interaction by a
cationic rhodacyanine dye analogue (MKT-077) resulted in nuclear translocation
and reactivation of p53 function sufficient to cause growth arrest of
transformed human cells (Wadhwa et al, 2000; 2002a). In tumors with wild type p53, the abrogation of mortalin-p53
interactions and reactivation of p53 function could be valid for cancer
therapy. Most common examples of these include breast carcinomas, glioblastomas
and teratocarcinomas.
Expression
of mortalin could be suppressed in malignant human fibroblasts using
specifically designed active hammerhead ribozymes. The cells with decreased
expression of mortalin undergo growth arrest and show reactivation of wild type
p53 function (Wadhwa et al, 2003b). However, the cells that lack p53 function also
experienced growth arrest suggesting that mortalin is involved in functions
other than p53 inactivation and are crucial for continued proliferation of
cancerous cells. One possibility could be due to its role as mitochondrial
importer as demonstrated in yeast with its homologue, SSC1p. The yeast homologue of mortalin, SSC1p, was shown
to be vital for mitochondrial import (Geissler et al, 2001;
Krimmer et al, 2000)
and its knock-out resulted in cell death (Craig et al, 1989). SSC1p was shown to bind to
Tim-44, an inner mitochondrial membrane anchor, and forms an essential
component of mitochondrial import machinery (Krimmer et al, 2000;
Strub et al, 2001).
Other proposed functions of SSC1p include unfolding of proteins outside
mitochondria, unidirectional translocation across mitochondrial membranes
initiated by membrane potential MĘY, completion of import by acting
as an ATP-driven motor and degradation of misfolded
peptides by m-AAA and PIM1 proteases in mitochondria (Lim et al, 2001; Liu et al, 2001). These functions may be critical for continued
proliferation of cancerous cells and thus targeting of mortalin may arrest the
growth of these cells.

Figure 2. Predictive
anti-apoptotic function of mortalin.
Its interaction with p53 in mitochondria may lead to abrogation of
p53-Bcl-2 association resulting in maintenance of anti-apoptotic functioning of
Bcl-2 protein.
Independent studies have assigned multiple functions to mortalin. They range from stress response (Craig et al, 1998; Merrick et al, 1995; Sadekova et al, 1997; Carette et al 2002; Resendez et al, 1986; Schneider and Hood, 2000; Wu et al, 1999), muscle activity