|Home | About | Journals | Submit | Contact Us | Français|
Since its original cloning by subtraction hybridization in 2002, it is now evident that Astrocyte elevated gene-1 (AEG-1) is a key contributor to the carcinogenic process in diverse organs. AEG-1 protein expression is elevated in advanced stages of many cancers, which correlates with poor survival. In specific cancers, such as breast and liver cancer, the AEG-1 gene itself is amplified further supporting a seminal role in tumorigenesis. Overexpression and inhibition studies both in in vitro and in in vivo models reveal the importance of AEG-1 in regulating multiple physiologically and pathologically relevant processes including proliferation, invasion, metastasis and gene expression. AEG-1 is a single-pass transmembrane protein with multiple nuclear localization signals and no known domains or motifs. Although pertinent roles of AEG-1 in the carcinogenic process are established, its potential function (promotion of metastasis only versus functioning as a bona fide oncogene) as well as localization (cell surface versus nucleus) remain areas requiring further clarification. The present review critically evaluates what is currently known about AEG-1 and provides new perspectives relative to this intriguing molecule that may provide a rational target for intervening in the cancer phenotype.
With completion of the sequencing of the human genome, it is now evident that we only know the precise functions of very few genes and there are a vast number of as-yet unexplored genes and their encoded proteins that might have profound roles in regulating key physiological and pathological events. This is particularly relevant for cancer in which novel genetic and epigenetic changes are being discovered every day. New candidate genes from previously unappreciated pathways, such as ubiquitination, metabolism, and mRNA biogenesis, are being identified that regulate the processes of tumorigenesis and cancer progression. This accrued knowledge is allowing researchers to develop strategies targeting novel molecules and unique aspects of carcinogenesis thereby ushering in optimism that prolonged survival in cancer patients will be an achievable objective. Astrocyte elevated gene-1 (AEG-1) represents an important genetic determinant regulating multiple events in tumorigenesis. Following its initial cloning in 2002 (1), AEG-1 has become the center of attention in an increasing spectrum of tumor indications for its multiple roles in regulating cancer progression and metastasis. The present review provides a current perspective of our understanding of this very interesting gene.
AEG-1 was first cloned in 2002 as a human immunodeficiency virus (HIV)-1- and tumor necrosis factor (TNF)-α-inducible gene in primary human fetal astrocytes (PHFA) (1). HIV-1 causes a low productive, non-cytolytic infection of astrocytes and does not infect neurons (2, 3). However, late stage acquired immunodeficiency syndrome (AIDS) is associated with neuronal degeneration and dementia (HIV-associated dementia or HAD) and it is hypothesized that changes in astrocyte function following HIV infection is a primary contributor to neuronal death (2). In an effort to elucidate gene expression changes in HIV-1-infected astrocytes, a rapid subtraction hybridization (RaSH) approach was performed between HIV-1-infected and non-infected PHFA (1). A series of HIV-1-induced (termed Astrocyte Elevated Genes or AEGs) and HIV-1-suppressed (termed Astrocyte Suppressed Genes or ASGs) genes were identified by this method (1). AEG-1 was shown to be induced by HIV-1, its cell surface protein gp120 or by HIV-1-induced cytokine TNF-α at 3 to 7 days following treatment indicating that AEG-1 induction by HIV-1 is a delayed and sustained event (1, 4). As yet, the role of AEG-1 induction in the alteration of astrocyte function or generation of HAD has not been elucidated. However, expression analysis revealed that AEG-1 expression is significantly elevated in cancer cells when compared to normal cells thereby highlighting novel and unanticipated functions of AEG-1 in tumorigenesis (discussed later).
In a separate approach in 2004, in vivo phage screening resulted in the cloning of mouse AEG-1 as a protein mediating metastasis of mouse breast cancer cells to lung and was named metadherin (5). A phage library, enriched for secreted and transmembrane proteins from mouse 4T1 breast cancer cells, were injected intravenously in mouse and phage that localized to the lungs were isolated and sequenced. One of the clones encoded the a.a. 378-440 of mouse AEG-1 which was designated as the ‘lung homing domain’ of the protein. In the same year, the mouse/rat AEG-1 was also cloned as a tight junction protein named LYRIC and also by gene trapping techniques and was named 3D3/lyric (6, 7). Human AEG-1 mRNA encodes a single-pass transmembrane protein of predicted molecular mass of ~64-kDa and pI of 9.3 (4). Human AEG-1 gene consists of 12 exons/11 introns and is located at chromosome 8q22 (4). This location is especially significant since amplification of 8q22 has been demonstrated in a number of malignancies such as malignant glioma, hepatocellular carcinoma (HCC) and breast cancer and indeed genomic amplification of AEG-1 has been identified in HCC and breast cancer patients (8–10). Human AEG-1 is a lysine-rich highly basic protein consisting of 582 a.a. residues (4, 11). It has three putative nuclear localization signals, between 79-91 a.a. residues, 432-451 a.a. residues and 561-580 a.a. residues (12). There is a putative transmembrane domain (TMD) between 51-72 a.a. residues. A lung homing domain has been identified in mouse AEG-1 (metadherin) that corresponds to 381-443 a.a. residues of human AEG-1 (5). Additionally, the molecule also contains an N-terminal “LXXLL” motif that is employed by transcriptional co-activators to interact with transcription factors (13).
Multi-tissue Northern blots containing total RNA of normal human organs revealed that AEG-1 mRNA is ubiquitously expressed at varying levels in all organs (4). Expression analysis in cell lines revealed that AEG-1 expression is significantly higher in breast, prostate, esophageal and liver cancer, and melanoma, malignant glioma and neuroblasoma cell lines in comparison to their normal counterparts (4, 9, 10, 14–16). These observations in cell lines have been confirmed in patient-derived tumor samples mainly by immunohistochemistry in tissue microarrays (TMA). In TMA containing 9 normal liver samples and 109 human HCC samples of different stages and grades, all 9 normal liver samples showed very low to undetectable expression of AEG-1 (9). Seven (7) out of the 109 human HCC samples showed no staining for AEG-1. Out of the 102 AEG-1 positive human HCC samples (93.58%) a statistically significant correlation between AEG-1 expression level and disease stage were obtained (9). Additionally, mining of existing Affymetrix gene expression data from 132 HCC patients revealed a statistically significant increase in AEG-1 mRNA expression in HCC patients compared to normal liver or cirrhotic patients (9). Strong AEG-1 staining was observed in 17 out of 31 samples of breast carcinoma patients while AEG-1 staining was absent in 18 out of 20 samples of normal breast tissue (5). A study from China analyzing 225 breast cancer patients showed increased expression of the protein in 44.5% of cases that correlated with the progression of the disease (17). Moreover, AEG-1 expression inversely correlated with patient survival. Interestingly, in this study, the metastatic tumors showed more nuclear staining of AEG-1 (17). In a separate study analyzing 170 breast cancer patients, 47% showed moderate to high level of AEG-1 expression and AEG-1 expression was significantly associated with a higher risk of metastasis indicating that AEG-1 might be a prognostic factor for this disease (10). However, in tissue sections of human breast cancer used in this study, AEG-1 protein was localized predominantly in the cytoplasm. A study using 20 benign prostatic hyperplasia (BPH) and 20 prostate cancer patient samples revealed strong positive AEG-1 expression in 80% of the prostate cancer patients compared to only 10% of the BPH cases (14). In a separate study using TMA (tumor microarray) of 143 prostate cancer and 63 BPH cases, increased AEG-1 staining was observed in the cancer patients compared to BPH patients (12). Interestingly, in this study more nuclear staining of AEG-1 was observed in BPH patients compared to prostate cancer patients and decreased nuclear staining of AEG-1 was associated with increased Gleason grade and lower survival rate (12). The authors hypothesized that nuclear AEG-1 might have a function in normal prostate epithelial cells that is lost during tumorigenesis. In a separate group of 11 patients with prostate cancer bone metastasis, 81.8% of prostate bone metastasis showed higher AEG-1 expression compared to normal bone (12). Analysis of 168 esophageal squamous cell carcinoma (ESCC) patients revealed high AEG-1 expression in 92.9% cases compared with normal esophageal tissue and AEG-1 expression inversely correlated with patient survival (15). Expression analysis using 2 normal peripheral nerve tissue and 10 neuroblastoma tissues revealed high AEG-1 staining in neuroblastoma while no staining was observed in the peripheral nerves (16). Our recent studies with 7 normal skin, 22 nevi, 15 radial growth phase (RGP) melanoma, 25 vertical growth phase (VGP) melanoma and 12 lymph node metastasis of melanoma revealed strong AEG-1 expression in VGP and metastasis when compared to the benign disease or RGP melanoma (manuscript in preparation). Interestingly, in addition to cytoplasmic staining, nuclear AEG-1 staining was observed in VGP and metastatic melanoma but not in normal melanocytes, nevi or RGP melanoma. Our recent findings in malignant glioma patients also reveal high AEG-1 expression in >90% of cases compared to the normal brains (manuscript in preparation).
Gain-of-function (overexpression) and loss-of-function (inhibition) studies are providing insights into the functional significance of AEG-1 in the process of tumorigenesis. Experimental approaches to define functions included proliferation (analyzed by MTT, colony formation assay and soft agar assay), migration, Matrigel invasion and in vivo tumorigenesis and metastasis assays. Overexpression of AEG-1 resulted in increased proliferation of human liver and esophageal cancer, malignant glioma, neuroblastoma and melanoma cells (manuscript in preparation) (9, 15, 16). Enhanced expression of AEG-1 also increased Matrigel invasive potential of HeLa, human HCC, neuroblastoma, malignant glioma and melanoma cells and cloned rat embryonic fibroblasts (CREF) (9, 13, 16, 18, 19). As a corollary, knockdown of AEG-1 inhibited proliferation of human prostate cancer, neuroblastoma and melanoma cells and induced apoptosis in prostate cancer and neuroblastoma cells (14, 16, 20). AEG-1 knockdown also inhibited invasion of human prostate cancer, neuroblastoma, malignant glioma and melanoma cell lines (unpublished data) (14, 20). Interestingly, neither overexpression nor inhibition altered any of these in vitro properties of human breast cancer cells (10), although our recent studies document that overexpression of AEG-1 does increase proliferation and invasion of human breast cancer cells and knockdown of AEG-1 reverses these effects (manuscript in preparation). The reasons for these discrepant findings remain to be determined. Overexpression of AEG-1 in normal human cells, such as primary normal human fetal astrocytes (PHFA), normal immortal melanocytes (FM516-SV) and normal cloned rat embryonic fibroblasts (CREF), protects them from serum starvation-induced apoptosis indicating that as an anti-apoptotic protein AEG-1 might function as an oncogene (21). AEG-1 can synergize with Ha-ras to augment the transformed phenotype in immortal SV40 T antigen expressing human melanocytes (FM516-SV) as well as in PHFA further supporting an oncogenic function of AEG-1 (4). With the exception of immortality, CREF do not exhibit transformation-associated properties (19). However, stable overexpression of AEG-1 in CREF results in morphological transformation with increased invasion and soft agar growth demonstrating that AEG-1 alone can function as an oncogene for rodent cells (Fig. 1) (19).
In in vivo assays, using nude mice xenograft models, overexpression of AEG-1 in human HCC cells resulted in highly aggressive, angiogenic and metastatic tumors while inhibition of AEG-1 abrogated subcutaneous tumor formation by human HCC, neuroblastoma and melanoma cells and intracerebral tumor formation by human glioma cells (unpublished data) (Fig. 2) (9, 16). Similarly, CREF-AEG-1 cells also generated highly aggressive and angiogenic tumors in nude mice (19). The tumor sections revealed augmented expression of specific angiogenesis molecules including angiopoietin-1 (Ang1), matrix metalloprotease (MMP)-2 and HIF1-α (19). Supporting the angiogenic functions of AEG-1, overexpression of AEG-1 augmented tube formation of human umbilical vein endothelial cells (HUVEC) while siRNA inhibition of AEG-1 blocked vascular endothelial growth factor (VEGF)-induced tube formation of HUVEC in Matrigel (19). In an in vivo chicken chorioallantoic membrane (CAM) assay, AEG-1 siRNA inhibited capillary neovascularization of H4 human glioma cells confirming the pro-angiogenic properties of AEG-1 (19). Overexpression of AEG-1 augmented in vivo metastasis of human breast cancer cells and HEK293T cells especially to the lungs, while inhibition of AEG-1 reversed this phenomenon for both human and mouse breast cancer cells (5, 10). Animals injected with AEG-1-overexpressing breast cancer cells had decreased survival rate while AEG-1 inhibition increased the survival rate (10). It has been shown that overexpression of AEG-1 increased adhesion of breast cancer cells to the endothelium, while inhibition of AEG-1 abrogated this phenomenon (10). This increased endothelial adhesion has been postulated to be the mechanism of increased metastasis by AEG-1. However, the AEG-1-interacting molecule on the endothelial cells has not been identified.
AEG-1 promotes tumorigenesis by modulating multiple signal transduction pathways and altering global gene expression changes (Figure 3 and Table 1). The first signaling pathway identified as being activated by AEG-1 was NF-κB (18). In HeLa cells and human malignant glioma cells, upon TNF-α treatment (which induces AEG-1 expression), AEG-1 translocates into the nucleus where it interacts with the p65 subunit of NF-κB and augments NF-κB-induced gene expression (13, 18). Although AEG-1 does not directly bind to DNA, upon TNF-α treatment it interacts with p65 and CBP on the IL-8 promoter increasing IL-8 transcription (13). IL-8, an NF-κB downstream gene, positively regulated angiogenesis and metastasis and inhibition of NF-κB abrogated AEG-1-induced augmentation of soft agar growth and Matrigel invasion by HeLa cells (18). Deletion mutant analysis showed that the N-terminal 71 amino acids that contain the transmembrane domain and the ‘LXXLL’ motif are important in mediating AEG-1-induced invasion, soft agar growth and NF-κB activation (13). However, the p65-interaction domain of AEG-1 was mapped to 101-205 a.a. of AEG-1 (13). The ‘LXXLL’ motif of AEG-1 might mediate its interaction with CBP which provides a crucial connection of the AEG-1-NF-κB complex to the basal transcriptional machinery. This hypothesis needs to be experimentally proven by site-directed mutagenesis studies. NF-κB activation by AEG-1 has also been recently documented in prostate and liver cancer cells (9, 14). Interestingly, in human promonocytic cells, LPS treatment resulted in induction of AEG-1 expression by NF-κB activation (22). On the other hand, AEG-1 itself activated NF-κB and inhibition of AEG-1 prevented LPS-induced production of proinflammatory cytokines, such as TNF-α and PGE2 (22). These findings suggest a potential immunological role of AEG-1 in bacterial infection, as well as a key role of AEG-1 in inflammatory mechanisms of cancer.
A second pathway modulated by AEG-1 is the PI3K/Akt pathway. Interestingly, this pathway is not only activated by AEG-1 but also plays a key role in regulating AEG-1 expression (23). AEG-1 expression is significantly augmented by Ha-ras. Ha-ras activates the PI3K/Akt pathway that leads to binding of the transcription factor c-Myc to the E-box element in the promoter region of AEG-1 and induces AEG-1 transcription (23). AEG-1 in turn activates the PI3K/Akt pathway, by a mechanism not yet elucidated, that plays an important role in providing protection from serum starvation-induced apoptosis of normal cells (21). While c-Myc induces AEG-1 transcription, AEG-1 also induces c-Myc expression and in neuroblastoma cells it induces N-Myc expression thereby amplifying the tumorigenic effect (16, 23). By activating Akt, AEG-1 downregulates pro-apoptotic Bad and p21 and upregulates MDM2 nullifying p53 function, thus exerting its anti-apoptotic effect (23). Inhibition of AEG-1 in prostate cancer cells downregulates Akt activation and leads to upregulation of forkhead box (FOXO)3a activity and p27 resulting in apoptosis (14). In esophageal cancer cells, activation of Akt by AEG-1 leads to upregulation of cyclin D1 and downregulation of p27 (15). The PI3/Akt signaling pathway has also been shown to regulate AEG-1-induced angiogenesis (19). Dominant negative inhibition of Akt abrogated AEG-1-induced tube formation by HUVECs, inhibited AEG-1-induced expression of angiogenic markers Tie-2 and HIF-1α in HUVECs and U87 human glioma cells and inhibited AEG-1-induced activation of the VEGF promoter (19).
In HCC cells, in addition to the NF-κB and Akt pathways, AEG-1 also activates the MAP kinase pathway, notably the MEK/ERK and p38 MAPK pathways and inhibition of either of these pathways result in abrogation of AEG-1-induced invasion of HCC cells (9). AP-1, an oncogenic transcription factor downstream of MEK/ERK, has been shown to be activated by AEG-1 in human prostate cancer cells (14).
Microarray analysis focused at identifying AEG-1 downstream genes resulted in identification of several important pathways and molecules contributing to carcinogenesis. In HCC cells, AEG-1 overexpression resulted in alteration of a plethora of genes involved in invasion, senescence, chemoresistance, angiogenesis and metastasis, all inexorably linked to HCC pathogenesis (9). AEG-1 activates Wnt/β-catenin signaling via ERK42/44 activation leading to β-catenin nuclear translocation and upregulates LEF-1/TCF-1, the ultimate executor of the Wnt pathway (9). Inhibition studies demonstrate that activation of Wnt signaling plays a key role in mediating AEG-1 function. In breast cancer cells, similar microarray approaches identified downregulation of two cell death-inducing genes, TRAIL and BINP3, by AEG-1 (10).
One important attribute of AEG-1 is its ability to confer resistance to chemotherapeutic agents in human HCC, breast cancer and neuroblastoma cell lines (9, 10, 20). Microarray analysis identified several key genes contributing to chemoresistance in HCC cells (9). AEG-1 induces the transcription factor LSF (Late SV40 Factor) that directly upregulates thymidylate synthase (TS) (24). 5-fluorouracil (5-FU) is converted intracellularly into its active metabolite FdUMP that inhibits TS thus reducing the thymidine pool and increasing the uracil pool leading to the inhibition of DNA synthesis. Additionally, AEG-1 induces the 5-FU catabolizing enzyme Dihydropyrimidine dehydrogenase (DPYD). The combined induction of LSF and DPYD by AEG-1 contributes to 5-FU resistance in HCC cells (24). In breast cancer cells, AEG-1 induces several genes related to chemoresistance, of which the aldehyde dehydrogenase 3 family, member A1 (ALDH3A1) and the hepatocyte growth factor receptor (Met) were shown to play a role in mediating resistance to doxorubicin, paclitaxel and 4-hydroxycyclophosphamide (4-HC) (10).
Recent studies have identified BCCIPα [BRCA-2 and CDKN1A (p21)-associated protein] as a potential interacting partner of AEG-1 (25). BCCIPα binds to p21 and enhances p21-mediated inhibition of Cdk2 kinase. Loss of BCCIP impairs G1/S checkpoint activation following DNA damage and in conjunction with BRCA2, BCCIP plays a role in homologous recombination repair of DNA damage and contributes to maintenance of chromosome stability. BCCIPα expression is downregulated in breast cancer and glioma cells and overexpression of BCCIPα inhibits cell growth. A yeast-two hybrid screen identified BCCIPα as a potential-interacting partner with AEG-1 and the N-terminal 169 a.a. residues of AEG-1 were shown to mediate this interaction (25). AEG-1 overexpression results in enhanced proteasomal degradation of BCCIPα. However, whether interaction of AEG-1 with BCCIPα followed by BCCIPα downregulation plays any role in AEG-1-mediated invasion or tumor progression was not studied.
There are some disputed issues regarding the potential localization of AEG-1 and its function. As mentioned earlier, AEG-1 was first cloned in 2002 as an HIV-1- and TNF-α-induced gene and subsequently cloned and reported by multiple groups in 2004. It was cloned as LYRIC by Britt et al. as a protein localized to tight junctions in polarized rat and human prostate epithelial cells and by immunofluorescence LYRIC showed co-localization with the tight junction protein ZO-1 (6). In the same study it was shown that although in normal rat hepatocytes LYRIC localized in the tight junctions, in rat hepatoma cells AS30D as well as in human 293T cells overexpressing LYRIC, the protein was localized in the perinuclear region (6). The perinulcear localization of the protein, cloned as 3d3/lyric, was also shown by Sutherland et al. (7). 3d3/lyric was detected in the endoplasmic reticulum (ER) and perinuclear region as well as in nucleus, especially nucleolus, in HeLa cells and human fibrosarcoma cells HT1080 (7). On the other hand, Brown and Ruoslahti identified Metadherin as a cell surface protein in mouse breast cancer cells and HEK293T cells overexpressing Metadherin (5). It was hypothesized that the large C-terminal domain of the protein is displayed outside of the cell and facilitates homing of breast cancer cells to the lung and promotes metastasis (5). In human HCC cells, in unpermeabilized condition AEG-1 protein is detected only on the cell surface while upon permeabilization it is detected both in the cytoplasm as well as in the nucleus (unpublished data). We also observe nuclear staining in human HCC samples (unpublishded data). In HeLa cells and malignant glioma cells TNF-α treatment results in nuclear translocation of AEG-1 (13, 18). A recent study narrowed down the nuclear localization signals (NLS) of AEG-1 by deletion and mutation studies (12). It was shown that extended sequences outside of the canonical NLS (exNLS) are important to determine AEG-1 localization. The exNLS-3 (546-582 a.a.) predominantly mediates nuclear localization while ex-NLS-1 (78-130 a.a.) regulates nucleolar localization (12). Although the putative molecular mass of AEG-1 is ~64 kDa, in Western blot analysis the predominant band migrates at ~75 kDa followed by an ~64 kDa band. It was shown that exNLS-2 (415-486 a.a.) of AEG-1 is monoubiquitinated resulting in the ~75 kDa band that retains AEG-1 in the cytoplasm (12). However, the functional properties of the generated constructs were not analyzed. Does a mutant that cannot localize to the nucleus or nucleolus have comparable functional activity as wild type AEG-1? Similarly, what is the biological behavior of a mutant that cannot be ubiquitinated?
Based on the in vitro and in vivo studies in animals and expression analysis using patient samples, it is eminently clear that AEG-1 plays a decisive role in the process of tumorigenesis in multiple organs. However, despite our expanded understanding of this interesting gene, a large number of unanswered questions and areas requiring clarification need to be addressed. One important question is whether AEG-1 only promotes metastasis, as suggested by two studies in breast cancer, or does it also regulate crucial events earlier in the carcinogenic process, such as immortalization, transformation and increased proliferation as suggested by multiple studies in other cancer indications as well as using immortal normal cells? Simply put, can AEG-1 function as a bona fide oncogene or does AEG-1 require an initial transforming event to allow tumor progression? The answer to this question might be obtained be analyzing whether AEG-1 overexpression alone might immortalize or transform primary normal cells and from AEG-1-overexpressing transgenic mouse models. The relationship between localization of AEG-1 and its function is also another area that requires resolution. In prostate cancer, decreased nuclear staining of AEG-1 correlates with poor prognosis, whereas in breast cancer, HCC and melanoma advanced disease is associated with increased AEG-1 nuclear staining. What is the role of nuclear AEG-1 in regulating the tumorigenic process? Apart from NF-κB, does AEG-1 function as a transcriptional co-activator for any other transcription factor(s)? What is the receptor for cell surface AEG-1 that allows it to home to the endothelium? Can cell surface AEG-1 be targeted by a radionucleotide-coupled antibody approach for diagnosis and treatment of metastatic disease? Is it possible to develop a small molecule inhibitor of AEG-1? Elucidation of the crystal structure of AEG-1 will facilitate identification of a small molecule inhibitor.
AEG-1 is a distinctive molecule that has no homologue and it has no known domains or motifs. Thus, identification of AEG-1 interacting proteins might provide clues about its mechanism of action. Although AEG-1 protein expression is increased in a significant percentage of patients, the gene itself is amplified in a small subset of patients and the relevance of this amplification to phenotype requires further investigation. This raises the obvious question as to what other mechanisms elevate AEG-1 expression in cancer cells?
As emphasized in this mini-review, AEG-1 is an intriguing molecule whose multiple functions continue to be elucidated. Over time and with increased investigation by multiple laboratories many of the questions we raise will be answered and a more precise understanding of the role of this gene product in normal and abnormal physiology will be forthcoming. Expanded research will also validate the potential effectiveness of AEG-1 as a target for intervening in cancer progression and metastasis.
The present studies were supported in part by National Institutes of Health grants P01 CA104177, R01 CA097318, R01 CA134721 and P01 NS31492, the National Foundation for Cancer Research (NFCR) and the Samuel Waxman Cancer Research Foundation (SWCRF) to PBF and awards from the Goldhirsh Foundation for Brain Cancer Research and the Dana Foundation to DS. DS is the Harrison Endowed Scholar in the VCU Massey Cancer Center. PBF holds the Thelma Newmeyer Corman Endowed Chair in Cancer Research in the VCU Massey Cancer Center and is an SWCRF Investigator.