Main

Body cavity effusions may be the presenting sign of malignancy or the first manifestation of recurrence and relapse. Tumor spread via survival and proliferation of tumor cells in body cavity fluid is an important route of metastasis in malignancies such as ovarian carcinomas, and a frequent cause of morbidity in many malignancies. A more complete understanding of the factors and pathways that permit survival and proliferation of malignant cells suspended in body cavity fluids could lead to more effective intervention and prevention of this route of tumor spread.

In general, tumor cells differ from their normal cells of origin in blunting of the apoptotic response to cell stressors.1 Metabolic adaptations that enhance survival such as induction of hypoxia-responsive genes and activation of autocrine growth-promoting pathways also occur frequently in cancer cells, and have been demonstrated in malignant effusions in particular.2, 3 An inter-related group of eight proteins, designated inhibitors of apoptosis (IAPs),4 XIAP(X-linked inhibitor of apoptosis), ILP-2, c-IAP1, c-IAP2, ML-IAP (livin), NAIP, survivin (TIAP), apollon (Bruce) are emerging as clinically important promoters of tumor cell survival and as suppressors of tumor cell response to apoptosis-inducing cancer therapies. IAPs prevent apoptosis by blocking the activation of specific caspases, thereby preventing caspase-mediated cell degradation. The ability of XIAP, reportedly the most potent of the IAPs, to inhibit caspases 3, 7 and 9 places it in a critical position to block both intrinsic and extrinsic apoptotic pathways.5 Elevated expression of XIAP could thus be an underpinning of both enhanced survivability of cancer cells and of resistance to apoptosis-inducing cancer therapy, leading to clinical relapse. Conversely, pharmacologic suppression of XIAP could restore sensitivity to tumor cells with XIAP-mediated resistance to tumor therapy.6 Two compounds with XIAP-inhibitory activity, phenoxodiol and flavopiridol, are already in clinical trials.7, 8, 9, 10

We describe below an immunocytochemical survey of XIAP expression in benign and malignant effusions and washes, and demonstrate XIAP positivity in a significant proportion of malignant effusions. The clinical and biologic implications are discussed.

Methods

Sections (4 μm) from 81 malignant and 35 benign ethanol-fixed, formalin postfixed paraffin-embedded cell block specimens from 82 pleural effusions, 22 ascites, 11 pelvic or peritoneal washes and one pericardial effusion were deparaffinized, exposed to 3% hydrogen peroxide to block endogenous peroxidase activity, incubated for antigen retrieval with 0.01 M citric acid at pH 6.0 and at 100°C × 5 min, followed by cooling for 20 min, then incubated with monoclonal anti-XIAP (#610763, BD Biosciences, San Jose, USA) diluted 1:250 in PBS with 5% albumin and 5% goat serum at 4°C × 72 h, and developed using EnVision-Plus reagents (Dako) and diaminobenzidine as chromagen. Only distinctly particulate cytoplasmic staining was scored as positive and graded as 1+ (weak), 2+ (moderate), and 3–4+ (strong). Primary sources of tumors were as follows: 13 ovary; 11 lung; 13 breast; 7 gastric; 7 endometrial; 5 mesothelial; 5 carcinomas of unknown primary; 3 fallopian tube; 4 colon; 3 pancreas; 1 gall bladder; 9 hematopoietic.

Results

The summary data are presented in Table 1.

Table 1 Summary of XIAP staining in malignant and benign body cavity effusions and washes
Full size table

Benign effusions (n=35) were XIAP-negative (Figure 1a), or showed nonspecific or focal-weak staining of mesothelial cells and macrophages in both noninflammatory specimens and specimens with acute or chronic inflammatory cells. Neutrophils and lymphocytes were characteristically negative. In rare benign effusions (two out of 35), staining of moderate intensity in a significant number of cells, mostly histiocytes (Figure 1b), possibly artifactual, was noted.

Figure 1
figure 1

XIAP immunstaining of benign effusions. (a) Absence of staining of mesothelial cells, histiocytes and inflammatory cells, benign effusion. (b) Moderate positive staining in significant number of histiocytes in a rare benign effusion.

Full size image

Unlike staining in benign effusions, XIAP positivity was frequently detected in malignant effusions. Positive staining was manifested by granular cytoplasmic staining that readily distinguished malignant from nonmalignant cell populations (Figure 2a–d). Staining positivity varied among different tumor types, with high prevalence in carcinomas of ovarian (13/13, 100%), pulmonary (9/11, 82%) and endometrial origin (6/7, 86%), and mesotheliomas (4/5, 80%). Positivity ranged from strong and/or diffuse to weak and/or focal. In contrast, XIAP staining was less prevalent in effusions of metastatic mammary carcinoma (6/13, 46%). Overall, 54 out of 81 malignant effusions (67%) displayed XIAP positivity.

Figure 2
figure 2

XIAP immunostaining of malignant effusions: (a) ovarian adenocarcinoma (note nonstaining of surrounding benign mesothelial cells or histiocytes), (b) lung carcinoma, (c) mesothelioma, (d) breast carcinoma.

Full size image

Among tumor types found less frequently in effusions or peritoneal washes, positive staining was noted in two out of three pancreatic adenocarcinomas; one of the three fallopian tube carcinomas; one of one gallbladder carcinoma and no colon carcinomas (0/4).

Nine hematopoietic malignancies (six non-Hodgkin's lymphomas, one CLL, one AML, and one multiple myeloma) were examined. The myeloma and two of the non-Hodgkin's lymphomas displayed a Golgi-like staining distribution in scattered to rare cells, the significance of which is uncertain; several tissue sections of myeloma were nonstaining, although a subset of lymphoma tissue sections was found to be XIAP-positive (unpublished data). A case of primary effusion lymphoma was nonstaining.

Five of the five carcinomas of unknown primary, including one with signet-ring morphology, were strongly and diffusely positive. Seven gastric carcinoma-induced effusions were examined, two from the same patient. This patient's earlier specimen was XIAP-negative; however, the repeat specimen, taken 50 days later, was XIAP-positive. Of note, the earlier specimen showed a significantly higher percentage (17%) of apoptotic-like degenerated cells compared to the later specimen (7%), possibly related to upregulated expression and apoptotic-inhibitory activity of XIAP (Figure 3a and b). A metastatic gastric carcinoma of intestinal type showed only equivocal-to-weak staining. Overall, four of the seven gastric carcinoma specimens were positively stained.

Figure 3
figure 3

XIAP in sequential ascites samples of metastatic gastric carcinoma. (a) XIAP is negative in tumor cells. Note apoptotic-appearing cells (arrow). (b) XIAP is positive in ascites 50 days later. Specimen had fewer apoptotic-appearing cells than in (a).

Full size image

Although the vast majority of benign or reactive mesothelial cells were nonstaining or displayed weak or nonspecific staining, four of five malignant mesothelioma specimens were positively stained, suggesting that strong XIAP staining in a cell population with mesothelial differentiation may signify malignancy.

Discussion

One feature that distinguishes many cancer cells from their normal counterparts is the aberrant activation of pathways that promote survival and/or suppress apoptosis. Normal mesothelial cells that exfoliate into body cavities presumably survive only transiently and fail to proliferate, analogous to the inability of nontransformed cells suspended in soft agar in vitro to achieve anchorage-independent growth. In contrast, tumor cells of many types adapt well to suspension in body cavity effusions, and survive as well as proliferate.

Contrasting with tumor cell growth in tissue, survival and proliferation of malignant cells in body cavity effusions occur despite the absence of an adjacent capillary oxygen source; absence of adhesion to stromal or endothelial cells or to acellular matrix or adhesion molecules, all of which may have paracrine effects on tumor cells; the frequent presence of low oxygen tension, high lactate levels, low glucose and acidic pH,11, 12 all environmental stressors that might minimize survival and proliferation and/or induce cell death in nontransformed cells. Given the considerable differences in microenvironment from tumor cells growing in tissue phase, it is not surprising that tumor cells from effusions may differ biologically from cells in the associated solid tumor.3

We have previously demonstrated one presumptive survival-promoting adaptation of malignant cells suspended in effusions—aberrant elevation of GLUT1, a hypoxia-inducible glucose transporter that occurs in a significant percentage of malignant effusions.2 GLUT1 elevation may be an adaptative response of malignancy to low oxygen tension, possibly leading to increased generation of ATP via glycolysis, a well-known feature of malignancy first reported by Warburg13 in the 1920s.

Other biologic changes that have been proposed as permitting the selective survival of malignant cells in effusions include induction of autocrine pathways.3

The present findings suggest that high expression of XIAP is another common feature of carcinoma cells in effusions that may play a role in promoting tumor cell survival. Only a few studies have provided evidence for reduced apoptosis in malignant effusions. Significantly enhanced levels of soluble Fas, a molecule that inhibits Fas receptor-mediated apoptosis, were reported in malignant effusions, compared to levels in tuberculous effusions and transudates. Lung cancer cell lines were also shown to synthesize soluble Fas.14 Conversely, an activator of death-receptor-mediated apoptosis, tumor-necrosis factor-related apoptosis-inducing ligand (TRAIL), was synthesized by macrophages from various malignant effusions; levels were reported to be higher in effusions with fewer tumor cells.15

Among tumors commonly associated with development of malignant effusions, strong XIAP staining was most prevalent in ovarian, pulmonary and endometrial carcinomas, and less prevalent in mammary carcinomas. A significant degree of XIAP immunostaining allows for ready distinction of malignant from benign and reactive cell populations in a high percentage of cases, the only exceptions being rare benign specimens in which diffuse, possibly artifactual positivity of various cell types was noted. Of interest, the consistently low expression of XIAP in benign mesothelial cells contrasts with XIAP positivity noted in four of five malignant mesothelioma specimens studied, suggesting a specific diagnostic application.

Effusions caused by malignant hematopoietic cells showed variable degrees of positivity, consistent with positivity in a subset of non-Hodgkin's lymphomas as well as positivity in Reed–Sternberg cells in Hodgkin's disease,16 our unpublished data. Unexplained is the nature of positive XIAP staining in a Golgi-like distribution.

Since benign cells are nonstaining in the great majority of cases, XIAP immunostaining may be an acceptable means of establishing or confirming the malignant nature of suspicious cell populations in effusion or wash samples.

Elevated XIAP expression has been linked to resistance to therapy in diffuse B-cell lymphoma17 and AML18 and to more aggressive clinical behavior or higher tumor grade in renal,19 uterine cervical squamous and endometrial carcinoma (unpublished data). The degree of XIAP staining of tumor cells may be a means of identifying effusions, and possibly their associated solid tumors, most resistant to apoptosis-inducing cancer therapies (ie, those with strong XIAP expression), and in the near future allow additional triaging to XIAP-blocking drugs presently being developed and clinically tested.9, 10