Human astrocytomas co-expressing Fas and Fas ligand also produce TGFbeta2 and Bcl-2

Fas and TRAIL 'death receptors' as initiators of inflammation: Implications for cancer

Fas (CD95/APO-1) and TRAIL (CD253, TNFSF10, APO2) are members of a subset of the TNF receptor superfamily known as 'death receptors'. To date, the overwhelming majority of studies on Fas and TRAIL (TNF-related apoptosis-inducing ligand) have explored the role of these receptors as initiators of apoptosis. However, sporadic reports also suggest that engagement of the Fas and TRAIL receptors can lead to other outcomes such as cytokine and chemokine production, cell proliferation, cell migration and differentiation. Indeed, although transformed cells frequently express Fas and TRAIL, most do not undergo apoptosis upon engagement of these receptors and significant effort has been devoted toward exploring how to sensitize such cells to the pro-apoptotic effects of 'death receptor' stimulation. Moreover, the expression of Fas and TRAIL receptors is greatly elevated in many cancer types such as hepatocellular carcinoma, renal carcinoma and ovarian cancer, suggesting that such tumors benefit from the expression of these receptors. Furthermore, several studies have shown that tumor proliferation, progression and invasion can be impaired through blocking or downregulation of Fas expression, but the mechanistic basis for these effects is largely unknown. Thus, the characterization of Fas and TRAIL as 'death receptors' is a gross oversimplification, especially in the context of cancer. It is becoming increasingly clear that 'death receptor' engagement can lead to outcomes, other than apoptosis, that become subverted by certain tumors to their benefit. Here we will discuss death-independent outcomes of Fas and TRAIL signaling and their implications for cancer.

Pine pollen inhibits cell apoptosis-related protein expression in the cerebral cortex of mice with arsenic poisoning

Previous studies have demonstrated that pine pollen can inhibit cerebral cortical cell apoptosis in mice with arsenic poisoning. The present study sought to detect the influence of pine pollen on apoptosis-related proteins. Immunohistochemistry, western blotting and enzyme-linked immunosorbent assays were used to measure the levels of apoptosis-related proteins in the cerebral cortex of mice with arsenic poisoning. Results indicated that pine pollen suppressed cell apoptosis in the cerebral cortex of arsenic-poisoned mice by reducing Bax, Bcl-2 protein expression and increasing p53 protein expression.

Targeting apoptosis pathways in glioblastoma

The treatment of glioblastoma remains a major challenge for clinicians since these highly aggressive brain tumors are relatively resistant towards radio- and chemotherapy. The pathways that control apoptosis are altered in glioblastoma cells leading to resistance towards apoptotic stimuli in general. In this review we describe the alterations affecting the p53 pathway, the BCL-2 protein family, the inhibitor of apoptosis proteins and several growth factor pathways involved in the regulation of programmed cell death and define possible targets for new therapies within these apoptotic pathways in glioblastomas. Moreover, we review strategies to target death receptor pathways, most notably to render the glioblastoma cells more susceptible towards this approach without enhancing toxicity in general. Most of the strategies targeting apoptosis in glioblastomas presented here are in a pre-clinical stage of development, however, they all share the ultimative goal to improve the outcome for glioblastoma patients.

Antisense therapeutics for tumor treatment: the TGF-beta2 inhibitor AP 12009 in clinical development against malignant tumors

Overexpression of the cytokine transforming growth factor-beta 2 (TGF-beta2) is a hallmark of various malignant tumors including pancreatic carcinoma, malignant glioma, metastasizing melanoma, and metastatic colorectal carcinoma. This is due to the pivotal role of TGF-beta2 as it regulates key mechanisms of tumor development, namely immunosuppression, metastasis, angiogenesis, and proliferation. The antisense technology is an innovative technique offering a targeted approach for the treatment of different highly aggressive tumors and other diseases. Antisense oligonucleotides are being developed to inhibit the production of disease-causing proteins at the molecular level. The immunotherapeutic approach with the phosphorothioate oligodeoxynucleotide AP 12009 for the treatment of malignant tumors is based on the specific inhibition of TGF-beta2. After providing preclinical proof of concept, the safety and efficacy of AP 12009 were assessed in clinical phase I/II open-label dose-escalation studies in recurrent or refractory high-grade glioma patients. Median survival time after recurrence exceeded the current literature data for chemotherapy. Currently, phase I/II study in advanced pancreatic carcinoma, metastatic melanoma, and metastatic colorectal carcinoma and a phase IIb study in recurrent or refractory high-grade glioma are ongoing. The preclinical as well as the clinical results implicate targeted TGF-beta2 suppression as a promising therapeutic approach for malignant tumor therapy.

c-FLIP expression in colorectal carcinomas: association with Fas/FasL expression and prognostic implications

AIMS

Disruption of apoptotic cell death has been implicated in tumour aggressiveness in colonic carcinogenesis. The Fas-Fas ligand (FasL) system is involved in the execution of apoptosis induced by the immune system. c-FLIP protein constitutes an inhibitor of Fas and other (TRAIL) death receptor-mediated apoptosis. The aim of this study was to investigate the simultaneous expression of Fas, FasL and c-FLIP in relation to standard clinicopathological parameters and patients' outcome in colorectal cancer.

METHODS AND RESULTS

Levels of Fas, FasL and c-FLIP protein expression were quantified immunohistochemically in paraffin-embedded tissues from 90 patients. Immunopositivity was detected for Fas, FasL and c-FLIP in 71%, 35.5% and 68.8% of cases, respectively. Concurrent expression of Fas/FasL was seen in 28 samples (31%), of which 24 (85.7%) also displayed c-FLIP positivity (P = 0.04). c-FLIP overexpression (> 10%) tended to prevail marginally in higher stage tumours (P = 0.09). Additionally, FasL and c-FLIP adversely affected survival on both univariate (P = 0.001 and P = 0.0024, respectively) and multivariate analysis [hazard ratio (HR) 3.491, P = 0.005 and HR 2.960, P = 0.036, respectively].

CONCLUSIONS

The frequent expression and coexpression of Fas, FasL and c-FLIP in colorectal carcinoma implicates c-FLIP as an inhibitor of the Fas-FasL-induced death pathway in these tumours. Moreover, c-FLIP conveys independent prognostic information in the presence of classical prognosticators.

Inhibition of TGF-β2 with AP 12009 in Recurrent Malignant Gliomas: From Preclinical to Phase I/II Studies

Transforming growth factor-beta2 (TGF-beta2) is known to suppress the immune response to cancer cells and plays a pivotal role in tumor progression by regulating key mechanisms including proliferation, metastasis, and angiogenesis. For targeted protein suppression the TGF-beta2-specific antisense oligodeoxynucleotide AP 12009 was developed. In vitro experiments have been performed to prove specificity and efficacy of the TGF-beta2 inhibitor AP 12009 employing patient-derived malignant glioma cells as well as peripheral blood mononuclear cells (PBMCs) from patients. Clinically, the antisense compound AP 12009 was assessed in three Phase I/II-studies for the treatment of patients with recurrent or refractory malignant (high-grade) glioma WHO grade III or IV. Although the study was not primarily designed as an efficacy evaluation, prolonged survival compared to literature data and response data were observed, which are very rarely seen in this tumor indication. Two patients experienced long-lasting complete tumor remissions. These results implicate targeted TGF-beta2-suppression using AP 12009 as a promising novel approach for malignant gliomas and other highly aggressive, TGF-beta-2-overexpressing tumors.

Molecular neuro-oncology and the development of targeted therapeutic strategies for brain tumors. Part 5: apoptosis and cell cycle

Brain tumors are a diverse group of malignancies that remain refractory to conventional treatment approaches. Molecular neuro-oncology has now begun to clarify the transformed phenotype of brain tumors and identify oncogenic pathways that might be amenable to targeted therapy. Abnormalities of the apoptotic and cell cycle signaling pathways are key molecular features of many brain tumors and are currently under evaluation for potential therapeutic intervention. The apoptosis pathway has numerous targets for molecular therapeutic development, including p53, Bax, Bcl-2, cFLIP, effector caspases, growth factor receptors, phosphatidylinositol-3-kinase, Akt and apoptosis inhibitors. Current molecular treatment approaches include antisense techniques, gene therapy and small-molecule modulators and inhibitors. Potential targets of the cell cycle pathway include the cyclins, cyclin-dependent kinases, p53, retinoblastoma, E2F and the cyclin-dependent kinase inhibitors. Developmental molecular therapeutics for this pathway include adenoviral and gene therapy, small-peptide cyclin-dependent kinase modulators, proteasomal inhibitors and small-molecule cyclin-dependent kinase inhibitors. Several of these recently identified agents have begun evaluation in clinical trials. Further development of targeted therapies designed to modulate apoptosis and the cell cycle, and evaluation of these new agents in clinical trials, will be needed to improve survival and quality of life for patients with brain tumors.

c-FLIP expression in bladder urothelial carcinomas: its role in resistance to Fas-mediated apoptosis and clinicopathologic correlations

OBJECTIVES

To investigate the incidence of Fas (exon 9) mutations and the expression of Fas, Fas-Fas ligand (FasL) system, and cellular FLICE-like inhibitory protein (c-FLIP) in relation to standard clinicopathologic parameters and patient outcome in bladder carcinoma. Disruption of apoptotic cell death has been implicated in tumor aggressiveness in bladder urothelial carcinomas. The FasL system is involved in the execution of apoptosis induced by the immune system. c-FLIP protein constitutes an important endogenous inhibitor of Fas and other death receptor-mediated apoptosis.

METHODS

The expression of Fas, FasL, and c-FLIP was quantified immunohistochemically in paraffin-embedded tissues from 53 patients for whom clinical information was available. DNA extracted from the same samples was screened for mutations in Fas exon 9 by single-strand conformation polymorphism and sequencing. The effect of Fas, FasL, and c-FLIP on clinical outcome was assessed by univariate and multivariate analyses.

RESULTS

Positive immunostaining was detected for Fas, FasL, and c-FLIP in 72%, 66%, and 81% of cases, respectively. Concurrent expression of Fas and FasL was seen in 27 samples (51%), of which 22 (81.5%) also displayed c-FLIP positivity. FasL and c-FLIP expression increased with advancing stage but was absent from normal urothelium. None of the 53 urothelial carcinoma samples analyzed showed evidence of mutations by polymerase chain reaction single-strand conformation polymorphism and direct sequencing. Survival analysis demonstrated that although both FasL and c-FLIP expression adversely affected survival, only c-FLIP remained statistically significant on multivariate analysis.

CONCLUSIONS

The frequent expression and coexpression of Fas, FasL, and c-FLIP in urothelial carcinomas implicates c-FLIP as an inhibitor of the Fas-FasL-induced death pathway in these tumors. Moreover, c-FLIP conveys independent prognostic information in the presence of classical prognosticators.

Protein expression of Fas, Fas ligand, Bcl-2 and TGFbeta2 and correlation with survival in initial and recurrent human gliomas

Several studies have recently demonstrated that human gliomas express Fas, Fas ligand (FasL), Bcl-2 and TGFbeta2 at some degree. These factors are considered to interact with apoptotic processes and to have immuno-reactive potential. Their role for tumor evasion from the immune surveillance is currently under examination. To date, there is only limited information about the definite expression patterns of these four factors in human gliomas, particularly in pilocytic astrocytoma (PA) and recurrent tumors. We analyzed 75 human gliomas for the immunohistochemical expression of Fas, FasL, Bcl-2, and TGFbeta2: (1) 25 PAs (WHO grade I), (2) 25 primary glioblastomas (WHO grade IV), and (3) 25 paired initial and recurrent glioblastomas (WHO grade IV), respectively. Co-expression of all four factors was present in the majority of specimens, i.e. in 72% (18/25) of PAs and 88% (47/50) of primary glioblastomas. Pilocytic astrocytomas showed significantly higher scores of TGFbeta2 expression (p < 0.05) and significantly lower Fas, Fas ligand and Bcl-2 scores (p < 0.05) than glioblastomas. There were no significant expression differences in initial versus recurrent glioblastoma specimens. Likewise, no significant correlation was observed between protein expression and clinical parameters, i.e. total survival time or progression free survival time, as documented by Kaplan-Meier method and log rank-test. In conclusion, Fas, FasL, Bcl-2 and TGFbeta2 are differently expressed in PAs versus glioblastomas. These factors, however, are not associated with patient prognosis. The broad co-expression of these factors may enable new therapeutic approaches in the future.

Fas ligand/Fas system in the brain: regulator of immune and apoptotic responses

Apoptosis, also known as programmed cell death, is the major type of cell death involved in normal development, regeneration, proliferation and pathologic degeneration in the central nervous system (CNS). The apoptotic process can be divided further into two pathways depending on the involvement of mitochondria and related biochemical cascades. The internal pathway of apoptosis is initiated by a variety of cytotoxic stimuli and mediated by the release of cytochrome c and subsequent activation of downstream caspases. The external pathway is mainly triggered by ligation of death receptors such as Fas, tumor necrosis factor (TNF)-related apoptosis inducing ligand-R1 (TRAIL-R1), TRAIL-R2 and TNFRp55, and mediated by direct activation of upstream caspases. The Fas-FasL system has been known as a prototypic inducer of extrinsic cell death responsible for cell-mediated cytotoxicity, peripheral immune regulation, immune privilege and "counterattack" of malignant tumor cells against the host immune system. Fas and FasL are expressed in the normal CNS, and expression increases in inflamed and degenerated brains. Like other specialized tissues such as the eye and testis, the Fas-FasL system is thought to be involved in immune suppressed status in the CNS. Expression of Fas and FasL is significantly elevated in a variety of the neurologic disorders, suggesting the possibility that this system may play roles in degenerative and inflammatory responses in the CNS. Therefore, the FasL-Fas system should be considered as a double-edged sword in the CNS: maintaining the immune suppressed status in normal brain and inducing neuronal cell death and inflammation in a variety of neurologic disorders.

Expression of the pro-apoptotic protein ARTS in astrocytic tumors

BACKGROUND

Apoptosis (i.e., programmed cell death) plays a major role in the development of astrocytic tumors, which are the most common tumors of the central nervous system. ARTS, a proapoptotic protein that is localized in the mitochondria, promotes apoptosis by functioning as an XIAP antagonist and a caspase activator.

METHODS

To investigate the role of ARTS in astrocytoma, the authors examined protein expression and apoptotic activity in 72 astrocytic tumors, which included low-grade astrocytomas, anaplastic astrocytomas, and glioblastomas.

RESULTS

Whereas normal astrocytes did not express the ARTS protein, astrocytoma cells strongly expressed ARTS, and the expression of this protein increased with increasing tumor grade. Furthermore, increased levels of ARTS were significantly associated with higher rates of apoptosis (as measured using the terminal deoxynucleotidyltransferase-mediated deoxyuridine triphosphate nick end-labeling [TUNEL] assay as well as an immunohistochemical staining assay for active caspase-3) in these tumors. Levels of two other apoptosis-related proteins, p53 and Bcl-2, also were examined using immunohistochemical methods; ARTS expression was found to be positively correlated with expression of the former and negatively correlated with expression of the latter, which is known to possess antiapoptotic activity.

CONCLUSIONS

The results of the current study suggest that ARTS levels reliably reflect the ability of cells to undergo apoptosis, which serves as a defense mechanism against the development and progression of astrocytoma. Furthermore, ARTS expression, when taken into consideration in combination with tumor grade, was the only independent predictor of survival identified in the current analysis. Thus, the authors conclude that ARTS may possess utility as a prognostic marker, as well as a therapeutic tool, for patients with astrocytoma.

Expression of a soluble transforming growth factor-beta (TGFbeta) receptor reduces tumorigenicity by regulating natural killer (NK) cell activity against 9L gliosarcoma in vivo

Immunotherapy of gliomas has been forwarded as an attractive alternative to standard therapeutic modalities. Numerous observations indicate some therapeutic efficacy with this approach, but it is not curative in most reports. It is well established that gliomas suppress immune reactivity via a number of mechanisms, including expression CD95 ligand (CD95L), which induces apoptosis of immune effector cells, and secretion of immunosuppressive factors such as transforming growth factor-beta (TGFbeta). It has been hypothesized that abrogation of production or function of TGFbeta would improve immune reactivity to gliomas. To investigate this in a fashion that is translatable into clinical practice, we utilized a retroviral vector encoding a truncated, soluble form of the Type II receptor for TGFbeta (TFGbeta sr) and expressed it in the rat 9L gliosarcoma line (9L-TGFbeta sr). We then determined whether expression of TGFbeta sr affected in vitro sensitivity of 9L to lysis by immune effector cells, whether expression of TGFbeta sr affected tumorigenesis of 9L in vivo, and whether TGFbeta sr affected expression of immunity to 9L. In these experiments, we determined that 9L-TGFbeta sr was more susceptible than sham transfected 9L (9L-neo) to lysis by natural killer (NK) cells. We also determined that subcutaneously implanted 9L-TGFbeta sr was less tumorigenic than 9L-neo in syngeneic rats. Similarly, survival was extended by approximately 40% in rats given intracranial 9L-TGFbeta sr compared to 9L-neo. Finally, we determined that elimination of CD161+ cells resulted in comparable growth of 9L-neo and 9L-TGFbeta sr in vivo, indicating that NK or NK-like cells were responsible for the anti-tumor effects in this model.

T-cell immune responses in the brain and their relevance for cerebral malignancies

In order that cellular immune responses afford protection without risk to sensitive normal tissue, they must be adapted to individual tissues of the body. Nowhere is this more critical than for the brain, where various passive and active mechanisms maintain a state of immune privilege that can limit high magnitude immune responses. Nevertheless, it is now clear that immune responses are induced to antigens in the brain, including those expressed by cerebral malignancies. We discuss hypotheses of how this can occur, although details such as which antigen presenting cells are involved remain to be clarified. Antitumor responses induced spontaneously are insufficient to eradicate malignant astrocytomas; many studies suggest that this can be explained by a combination of low level immune response induction and tumor mediated immunosuppression. A clinical objective currently pursued is to use immunotherapy to ameliorate antitumour immunity. This will necessitate a high level immune response to ensure sufficient effector cells reach the tumor bed, focused cytotoxicity to eradicate malignant cells with little collateral damage to critical normal cells, and minimal inflammation. To achieve these aims, priority should be given to identifying more target antigens in astrocytoma and defining those cells present in the brain parenchyma that are essential to maintain antitumour effector function without exacerbating inflammation. If we are armed with better understanding of immune interactions with brain tumor cells, we can realistically envisage that immunotherapy will one day offer hope to patients with currently untreatable neoplastic diseases of the CNS.

Apoptosis and survival in high-grade astrocytomas as related to tumor Fas (APO-1/CD95) expression

Although a majority of high-grade gliomas express the apoptosis-inducing receptor Fas, little is known about the extent of apoptosis or prognostic significance of Fas expression in these tumors. In situ labeling of apoptotic cells and Ki-67 immunohistochemistry were performed on 51 high-grade human astrocytomas previously characterized for Fas expression. Survival data was compiled from patient records and correlated with tumor grade, apoptotic index (AI) and Fas expression. A significant correlation was found between tumor grade and the AI and Ki-67 labeling index (LI); however, only the AI increased significantly with Fas expression. The AI increased from 0.39 +/- 0.12% to 0.82 +/- 0.10% in grade III vs. IV astrocytomas (P = 0.003). The Ki-67-LI increased from 3.64 +/- 1.5% to 11.35 +/- 2.1% in grade III vs. IV astrocytomas (P = 0.004). Additionally, tumors expressing higher Fas levels had a greater AI than those expressing lower levels (0.81 +/- 0.11% vs. 0.43 +/- 0.11%) (P = 0.017). Despite longer median survivals for patients with tumors exhibiting high Fas expression, statistical significance was not achieved. Patients with grade III astrocytomas demonstrated a median survival of 20 vs. 18 months for tumors with high vs. low Fas expression (P = 0.51). Patients with grade IV astrocytomas demonstrated a median survival of 9 vs. 7.4 months for tumors with high vs. low Fas expression, respectively (P = 0.77). Although the degree of Fas expression in high-grade astrocytomas appears to correlate with the apoptotic rate, no overall differences in survival could be demonstrated between tumors expressing high vs. low Fas levels.

Possible involvement of Fas system in the induction of apoptosis in human astrocytic brain tumors

1. For a better understanding of the biological features of astrocytic tumors, we investigated apoptosis and its pathway, especially in the interaction between Fas and Fas ligand (FasL). 2. We examined the presence of apoptosis in human astrocytic brain tumors by terminal deoxynucleotidyl transferase (TdT)-mediated d-UTP-biotin nick end labeling (TUNEL) and then apoptotic index (AI) was calculated. We also examined the distribution of Fas and FasL-positive tumor cells immunohistochemically. Labeling index (LI) for Fas and FasL was calculated as Fas-LI and FasL-LI, respectively, and compared to AI. 3. Tumor cells expressing both Fas and FasL were TUNEL positive. Such cells were distributed sparsely in low-grade astrocytomas, but focally in glioblastomas. There was a close correlation among AI, Fas-LI, and FasL-LI, and astrocytic tumors with higher Al were associated with a longer survival time than that with lower Al. 4. It was concluded that the Fas system may be involved in the apoptosis of astrocytic tumors, and Al can be a useful parameter for assessing prognosis of astrocytic tumors.

Apoptosis of T lymphocytes invading glioblastomas multiforme: a possible tumor defense mechanism

OBJECT

The goal of this study was to investigate whether apoptosis occurs in T lymphocytes that invade Fas ligand (FasL)-expressing glioblastomas multiforme (GBMs) and if its induction could be mediated by Fas.

METHODS

Apoptotic T lymphocytes were detected in GBMs by using detection of cell-type markers combined with active caspase-3 immunohistochemical analysis, a recently introduced apoptosis-specific in situ ligation assay, as well as by examining morphological criteria. Apoptotic T cells expressed Fas and were localized in the vicinity or in direct contact with FasL-expressing tumor cells. The T lymphocytes were undergoing apoptosis in spite of Bcl-2 expression. Expression of Bax was also detected in dying T cells, which can explain the absence of the protective effect of Bcl-2. because Bax inhibits Bcl-2 death-repressor activity.

CONCLUSIONS

On the basis of the data presented in this paper, the authors suggest that GBM cells that express FasL can induce apoptosis in invading immune cells. This phenomenon may play an important role in these tumors' maintenance of immune privilege and evasion of immune attacks. Awareness of this phenomenon should be helpful for the development of novel strategies for treatment of malignant gliomas.

Tumor Fas (APO-1/CD95) up-regulation results in increased apoptosis and survival times for rats with intracranial malignant gliomas

OBJECTIVE

The cellular "death" receptor Fas has been proposed to be a potential specific target for anti-glioma therapy. However, little is known regarding the effects of Fas expression on glioma viability in vivo. The goal of this study was to clarify the relationships among Fas expression, apoptosis, and survival rates for high-grade astrocytomas.

METHODS

Fas expression was measured in several human glioblastoma multiforme cell lines and a malignant rat glioma cell line (36B10), before and after Fas up-regulation by gene transfer. Expression was correlated with the degree of Fas-mediated cytotoxicity and apoptosis induced after Fas activation. Subsequently, rats underwent intracranial implantation of either wild-type or genetically altered 36B10 cell lines, for study of the effects of Fas up-regulation on survival rates.

RESULTS

Low levels of cell surface Fas expression in glioblastomas multiforme were correlated with their limited susceptibility to Fas-mediated cytotoxicity. Through Fas receptor up-regulation, relationships among increased Fas expression, Fas-mediated cytotoxicity, and apoptosis were demonstrated. The percentage of cells undergoing apoptosis after exposure to a Fas ligand-producing cell line increased from 4% in the sham-transfected line (36B10-) to 27% in the Fas-transfected line (36B10-Fas). After intracranial implantation of these tumors into rats, the median survival time increased significantly from 14 days (36B10 and 36B10-) to 24.5 days (36B10-Fas), which represents a 75% increase in the survival time for the greater Fas-expressing group (P = 0.0005).

CONCLUSION

It seems that the overall low rate of apoptosis in high-grade astrocytomas is related to low levels of cell surface Fas expression. With increases in cellular Fas expression, rates of Fas-mediated apoptosis and survival rates were increased.

Soluble Fas-ligand (sFasL) in human astrocytoma cyst fluid is cytotoxic to T-cells: another potential means of immune evasion

Gliomas of all grades have been shown to express FasL, an apoptosis-inducing protein. Because of the ability of FasL to be cleaved from cell surfaces by metalloproteinases, soluble FasL can be released by FasL bearing cells into surrounding tissues. In the present study, we demonstrate the presence of sFasL in the cyst fluids of astrocytomas. Additionally, a human T-cell line, Jurkat, exposed to astrocytoma cyst fluid resulted in significantly increased cytotoxicity as compared to controls, an effect blocked by FasL neutralizing antibodies. This suggests that sFasL, may be utilized as a means of escaping immune surveillance by these tumors.