Ann Thorac Surg 2002;73:1567-1571
© 2002 The Society of Thoracic Surgeons
Original article: general thoracic
Cell membrane fluidity and prognosis of lung cancer
Miha Sok, MD, PhD*a,
Marjeta 
entjurc, PhDb,
Milan Schara, PhDb,
Janez Stare, PhDc,
Toma
Rott, MD, PhDd
a Department of Thoracic Surgery, University Medical Centre Ljubljana, Slovenia
b J. Stefan Institute Ljubljana, Slovenia
c Institute of Biomedical Informatics, Medical Faculty Ljubljana, Slovenia
d Institute of Pathology, Medical Faculty, Ljubljana, Slovenia
Accepted for publication December 30, 2001.
* Address reprint requests to Dr Sok, University Medical Centre Ljubljana, Department of Thoracic Surgery, Zalo
ka 7, 1000 Ljubljana, Slovenia
e-mail: miha.sok{at}mf.uni-lj.si
 |
Abstract
|
|---|
Background. Membranes of tumor cells have been found to posses higher fluidity than membranes of nontumor cells. Plasma membrane fluidity is significantly correlated with malignant potential of these cells.
Methods. Seventy-five patients operated on for lung cancer were studied prospectively. During the operation, lung tumor samples were taken from the resected lung for evaluation by electron paramagnetic resonance. The fluidity variable H13, which is proportional to the plasma membrane fluidity, was determined from the electron paramagnetic resonance spectra. The association between H13 and survival was determined by survival analysis using Kaplan-Meier curves and Cox regression.
Results. Pathologic TNM stage and the fluidity variable H13 were the only prognostic variables significantly associated with survival time in multivariate proportional hazards regression model. Thus, H13 was shown to be an independent prognostic variable for survival, which was also confirmed by a separate analysis relating the TNM stage and H13. Dividing the patients into two groups, one with an H13 value higher than the median and another with H13 below the median, resulted in significantly different survival curves (p = 0.01).
Conclusions. Patients with high plasma membrane fluidity, indicated by high H13 of the resected lung tumor tissue, seem to have poorer prognosis than those with less fluid membranes. We suggest that the fluidity variable could be used as an independent additional prognostic factor and a tool to identify patients who may be helped by adjuvant postoperative therapy.
|
Introduction
|
|---|
Membrane fluidity is a characteristic of cell membranes that depends on their lipid and protein composition. Plasma membrane fluidity of cancer cells appears to be significantly correlated with the malignant potential of these cells [1–4]. Evidence is accumulating that alterations in the plasma membrane fluidity of cancer cells are closely correlated with the capability of cancer cells to form metastases [2, 5]. According to recent investigations, the plasma membrane is a mosaic of domains possessing different degrees of fluidity. This heterogeneity may be relevant in the context of metastatic dissemination of tumors [6]. Fluidity is an important characteristic of biologic membranes at a molecular level. It indicates the way and the rate of motion of molecules in the membrane, and is inversely related to membrane microviscosity. Changes in membrane fluidity may play a role in the regulation of membrane properties, both under physiologic conditions and in the pathogenesis of the disease. Knowing the characteristics of membrane fluidity is therefore essential to understanding the complex mechanisms regulated by membrane properties [7]. Changes in the plasma membrane fluidity of tumor cells may affect antigens and receptors [8, 9], as well as the cancer cell motility and capacity to infiltrate the basement membrane [10] and the deformability potential of metastatic cells [5].
Our previous electron paramagnetic resonance (EPR) study of resected lung tissues showed that more fluid membranes and membrane domains were present in the tumor tissue than in normal lung tissues .The observation accords with some previous reports on tumor cell cultures, which usually showed higher membrane fluidity than their normal analogs [10–12]. The fluidity variable H13 was therefore defined as a simple criterion that may be used to correlate EPR and clinical data [13]. It reflects the relative ratio between the most fluid and the least fluid membrane domains in the lung cancer tissue. The most important prognostic indicator in patients with non–small-cell lung cancer is the pathologic (p-) stage of the disease [14–16]. The TNM classification, however, does not always provide objective information on the p-stage of the disease. Routine microscopic examination of the resected lung cancer and lymph nodes, used as a sine qua non for the TNM staging, has some limitations. Thus, additional immunohistochemical studies revealed lung cancer micrometastases in more than 60% of lymph nodes that had been negative on conventional examinations, suggesting an occult and undetected metastatic disease and down-staging of the tumor [17]. The p-stage largely depends on the stage of the tumor development at which the diagnosis was made, and is not necessarily related to the malignant potential of cancer cells. It is a function of time rather than of malignant potential. Many other cellular markers have been studied for their prognostic value in lung cancer after resection [18, 19].
The fluidity variable H13, in contrast to p-stage, provides information about the malignancy of cells in lung tumor tissue independent of the time of diagnosis. The purpose of this study was to determine whether membrane fluidity, indicated by the empirical variable H13, can be used as an additional prognostic variable in lung cancer.
|
Patients and methods
|
|---|
Patients
We studied a random series of 75 lung cancer patients, operated on at the Department of Thoracic Surgery, University Medical Center Ljubljana, from June 1988 to June 1995. There were 10 women and 65 men; their average age was 61 years. The diagnosis was based on bronchoscopy, biopsy, peribronchial needle biopsy, pulmonary function tests, and computerized tomography, performed in another institution. In all patients, the diagnosis of non–small-cell lung cancer stages I and II was verified cytologically and histologically. There were 29 of 75 (39%) epidermoid carcinomas, 31 of 75 (41%) adenocarcinomas, and 15 of 75 (20%) large cell carcinomas. The operative procedures included 43 of 75 (57%) lobectomies, 23 of 75 (31%) pulmonectomies, and 9 of 75 (12%) nontypical resections. An R0 resection was invariably performed in all patients. The resected samples were examined by the same pathologist at the Institute of Pathology, Medical Faculty, Ljubljana, and then classified by the TNM system. There were 37 (49%) patients with p-stage I disease, 8 (11%) patients with p-stage II, 25 (33%) patients with p-stage III, and 5 (7%) patients with p-stage IV disease. Patients classified as p-stages III and IV were postoperatively irradiated, and were treated as a single group in further analysis.
Electron paramagnetic resonance measurements
During the operation, lung cancer tissue samples from the tumor periphery and normal pulmonary tissue samples were obtained for EPR measurements in all patients. The measurements were performed at the J. Stefan Institute, Ljubljana within 2 hours of the operation. The samples were cut into slices, approximately 0.5 mm thick, weighing 10 to 20 mg, and placed at room temperature for 30 minutes into 0.1 mmol/L solution of lipophilic spin probe methylester of 5-doxyl palmitate (MeFASL) [16]. The spectra were recorded on a Varian E-9 X-band spectrometer: microwave power, 20 mW; modulation frequency, 100 kHz; modulation amplitude, 0.2 mT at room temperature. From the EPR spectra an empirical variable H13 (H13=h3/h1x10), the fluidity variable, was calculated [13]. The mean (± standard deviation [SD]) fluidity variable H13 in the tumor tissue was 24.1 ± 11.7, and in normal lung tissue, 10 ± 0.4. The distribution of H13 in tumor tissue was notably skewed to the right, and the median H13 of 21 was considered a more appropriate measure of location. The first and the third quartiles were 17 and 28, respectively.
Follow-up
The patients were followed up for at least 60 months; at 3-month intervals during the first 2 years, at 6-month intervals for the next 3 years, and yearly thereafter. They were evaluated by physical examination, chest radiography, and by computed tomographic scanning or ultrasound, as clinically indicated. The date of the first relapse and the date of death—for nonsurvivors—were recorded. The postoperative lung cancer–specific mortality was calculated. Four of the 75 patients succumbed to noncancer diseases and showed no signs of lung cancer at autopsy. They were included in the survival statistics for the period until their death.
The prognostic value of H13 was determined in a group with H13 values less than or equal to the median (H13 
21), and in a group with H13 values greater than the median (H13 > 21).
Statistics
The Cox proportional hazards model was used to study the relationship between H13 and survival time after the operation. The analysis was performed using the library Design programs [20], now distributed with the S-Plus 2000 statistical software [21]. Restricted cubic splines were used to model the dependency of the hazard function on H13. The hazard function gives the instantaneous potential per unit time for the event to occur, given that the individual has survived up to the time t.
|
Results
|
|---|
Table 1
summarizes clinical variables of two study groups. The groups with different H13 levels showed similar demographic variables.
Multivariate survival analysis using the Cox proportional hazards model showed that the p-stage and the H13 fluidity variable were prognostic indicators of postoperative survival after lung cancer resection in contrast to sex, age, histologic type of tumor, and type of operation, which showed no prognostic value. The results of the Cox model fit for the two prognostic indicators are presented in Table 2.
The p-stage was included in the model to adjust for possible confounding. Inasmuch as a variable stage has three values (TNM stages III and IV were treated together), two dummy variables—stages II and III—were created. The fit was not significantly improved by any other variable, such as age, sex, histology, and type of operation. No evidence was found for the interaction between H13 and the p-stage, implying that H13 does not change significantly with different stages. For p-stage I, II, and III the mean values were 22.7 ± 10.7, 20.8 ± 9.0, and 27.3 (± 12.5, respectively. As indicated by the dot plot in Figure 1,
one-way analysis of variance revealed no statistically significant differences in H13 values among different stage groups (p = 0.18), implying that H13 is an independent prognostic indicator.
View this table:
[in this window]
[in a new window]
|
Table 2. Estimated Coefficients, Standard Error, and p Values for Proportional Hazards Model Containing the Fluidity Variable H13 (modeled with restricted cubic splines) and Stages of Diseasea
|
|
Restricted cubic splines [19] were used to show either a linear or nonlinear relationship between H13 and survival. The standard use of the Cox model assumes—sometimes unjustifiably—that the logarithm of the hazard function is linearly related to prognostic variables. We used restricted cubic splines [22] to indicate a possible nonlinear relationship; H13', H13'', H13''' in Table 2 are the components of the spline function. Figure 2
shows the predicted log relative hazard (relative to the reference H13 value at the median) as a function of H13. Clearly, the relation is not linear throughout the H13 range. Figure 2 shows nonlinearity at H13 of 35 or less. H13 values below 35 are optimal to survival, whereas H13 values greater than 35 account for poor prognosis.
View larger version (7K):
[in this window]
[in a new window]
|
Fig 2. Predicted log relative risk (relative to the reference value at the median, for tumors H13 = 21) as a function of H13, assessed by Cox modeling with restricted cubic splines.
|
|
The value that is usually reported with the Cox model is the relative hazard for two groups differing by 1. This makes sense with the stage, but not with H13, for which the relations are further complicated because of nonlinearity. We therefore report relative hazards for H13 values differing by 10. The results are presented in Table 3.
The survival curves for groups with different p-stages are plotted in Figure 3A.
Figure 3B shows Kaplan-Meier survival curves for the group with H13 less than or equal to 21, and the group with H13 more than 21. The probability values for the test of equality of the survival curves are p-stage (p < 0.0001) and H13 variable (p < 0.01).
View larger version (9K):
[in this window]
[in a new window]
|
Fig 3. (A) Five-year survival curves in lung cancer patients by the TNM stage (p < 0.0001). (B) Five-year survival curves for the group with H13 less than or equal to the median (H13 21), and the group with H13 greater than the median (H13 > 21; p < 0.01).
|
|
|
Comment
|
|---|
The fluidity variable of human lung cancer tissue reflects the physiologic state of lung cancer cells. Our previous work showed that the membranes of tumor cells are more fluid than those of normal lung cells [13]. This observation agrees with the results of other studies that reported increased fluidity of plasma membrane in cancer cell cultures, and stressed the correlation between higher fluidity and metastatic potential of cancer cells [1–4]. The malignant potential is presented by different cellular markers influencing postoperative survival [18, 19]. Neuroendocrine differentiation of tumor tissue in patients with small cell and non–small-cell lung cancer has been recently found to be associated with a high malignant potential and poorer prognosis [23]. In our study, lung tumor tissue was not tested for neuroendocrine differentiation or other cellular markers. Further studies will be needed to determine possible correlation between these markers and plasma membrane fluidity.
Our previous study [13] also showed that prognostic criteria usually used to define the malignant potential of tumor cells, such as tumor histology, quantitative presence of different tumor and nontumor cells, and the pathologic stage of the disease, had no significant influence on fluidity. On the other hand, it is known that the degree of tumor cell differentiation is of limited prognostic value in patients with lung cancer [24], and that plasma membrane fluidity of cells in tumor tissue affects the surrounding cells [25, 26]. It is not surprising that the above-mentioned prognostic criteria are not necessarily correlated with the fluidity variable H13 measured in tumor tissue. Similar results were obtained for polysialic acid, in which positive expression in patients with resected non–small-cell lung cancer irrespective of the disease stage was found to account for a significantly poorer prognosis [18]. The TNM classification, considered the leading prognostic tool in patients with non–small-cell lung cancer [14–16], does not always provide objective information about the stage of the disease. It indicates the timing of tumor development and diagnosis, but is not necessarily related to the malignant potential of cancer cells. On the other hand, the fluidity variable H13 reflects the status of tumor cells, and depends primarily on the cholesterol-to-phospholipid ratio in the membrane, on the degree of unsaturation of the phospholipid acyl chains in the bilayer, and on the amount and distribution of proteins [27]. There are many different factors that can influence the fluidity of plasma membranes in malignant cells. The exact reason for the decrease in plasma membrane fluidity is not yet known, but different factors are implicated. One important reason for decreased fluidity is reduced cholesterol content, typical for tumor cells. However, tumor necrosis factor, some fatty acid esterification enzymes, or some other proteins might also exert some biologic influence on the dynamic properties of membranes in tumor cells [1, 26]. The major drawback of the fluidity variable measurement is that it indicates the stage of very heterogeneous population of cells in lung tumor tissue. The fluidity variable reflects the ratio between the tissue regions with greater and lesser fluidity. Because it does not provide accurate information on the status of malignant cells, but rather gives an overall estimate of all cells in the tissue, it cannot be used as the only variable in the prognosis of cancer. Yet, a careful analysis of the results of our study, which took into consideration the survival of patients followed up for 5 years after operation, as well as the EPR measurements, confirmed a positive correlation between H13 and survival. The H13 value provides information on the malignancy of tumor cells at the molecular level, which does not always accord with the p-stage of the disease. Our results showed that in cancer patients, H13 can be used as an independent prognostic variable that does not necessarily correlate with the p-stage of the disease (Fig 2). For example: in 8 patients with p-stage III disease and H13 of less than 26 (range, 10 to 26; mean, 17.2), survival was more than 36 months (range, 36 to 60 months; mean, 50.6 months), and in 3 p-stage I patients with H13 of 31 to 38 (mean, 34.3), it was less than 17 months (mean, 9.6 months). In these cases the prognosis correlated more strongly with the H13 values than with the p-stage.
In conclusion, the plasma membrane fluidity variable H13 determined in resected human lung cancer samples has proved a valuable postoperative prognostic factor, providing useful independent information about the patient’s disease. It may have an important role in facilitating therapeutic decisions after operation. Further studies will be needed to identify patients or patient groups with early-stage disease whose risk of recurrence, determined by the H13 criteria, is sufficiently high to justify adjuvant therapy.
|
Footnotes
|
|---|
This work was supported by the Ministry of Education, Science and Sport, research project J3/0355.
|
References
|
|---|
-
Shinitzky M. Membrane fluidity in malignancy. Adversative and recuperative. Biochim Biophys Acta 1984;738:251-261.[Medline]
-
Deliconstantinos G. Physiological aspects of membrane lipid fluidity in malignancy. Anticancer Res 1987;7:1011-1021.[Medline]
-
Kier A.B. Membrane properties of metastatic and non-metastatic cells cultured from C3H mice injected with LM fibroblast. Biochim Biophys Acta 1990;1022:365-372.[Medline]
-
Kozlowska K., Nowak J., Kwiatkowski B., Cichorek M. ESR study of plasmatic membrane on transplantable melanoma cells in relation to their biological properties. Exp Toxicol Pathol 1999;51:89-92.[Medline]
-
Nakazawa I., Iwaizumi M. A role of the cancer cell membrane fluidity in the cancer metastases: an ESR study. Tohoku J Exp Med 1989;157:193-198.[Medline]
-
Sherbet G.V. Membrane fluidity and cancer metastasis. Exp Cell Biol 1989;57:198-205.[Medline]
-
Bloom M. The physics of soft, natural materials. Phys Can 1992;48:7-16.
-
Daefler S., Krueger G.R., Modder B., Deliconstantinos G. Cell membrane fluidity in chronic lymphocytic leukemia (CLL) lymphocytes and its relation to membrane receptor expression. J Exp Pathol 1987;3:147-154.[Medline]
-
Iwagaki H., Marutaka M., Nezu M., Suguri T., Tanaka N., Orita K. Cell membrane fluidity in K562 cells and its relation to receptor expression. Res Commun Mol Pathol Pharmacol 1994;85:141-149.[Medline]
-
Taraboletti G., Perin L., Bottazzi B., Mantovani A., Giavazzi R., Salmona M. Membrane fluidity affects tumor cell motility, invasion and lung colonizing potential. Int J Cancer 1989;44:707-713.[Medline]
-
Inbar M. Fluidity of membrane lipids: a single cell analysis of mouse normal lymphocytes and malignant lymphoma cells. FEBS Lett 1976;67:180-185.[Medline]
-
Hattori T., Andoh T., Sakai N., et al. Membrane phospholipid composition and membrane fluidity of human brain tumour: a spin label study. Neurol Res 1987;9:38-43.[Medline]
-
Sok M., entjurc M., Schara M. Membrane fluidity characteristics of human lung cancer. Cancer Lett 1999;139:215-220.[Medline]
-
Kirsh M.M., Rothman H., Argenta L., et al. Carcinoma of the lung; results of treatment over ten years. Ann Thorac Surg 1976;21:371-377.[Abstract]
-
Naruke T., Goya T., Tsuchiya R., Suemasu K. Prognosis and survival in resected lung carcinoma based on the new international staging system. J Thorac Cardiovasc Surg 1988;96:440-447.[Abstract]
-
Capwell S., Sudlow M.F. Performance and prognosis in patients with lung cancer. The Edinburgh Lung Cancer Group. Thorax 1990;45:951-956.[Abstract/Free Full Text]
-
Chen Z.L., Perez S., Holmes E.C., et al. Frequency and distribution of occult micrometastases in lymph nodes of patients with non–small-cell lung carcinoma. J Natl Cancer Inst 1993;85:493-498.[Abstract/Free Full Text]
-
Tanaka F., Otake Y., Nakagawa T., et al. Prognostic significance of polysialic acid expression in resected non–small cell lung cancer. Cancer Res 2001;61:1666-1670.[Abstract/Free Full Text]
-
Pelletier M.P., Edwardes M.D., Michel R.P., Halwani F., Morin J.E. Prognostic markers in respectable non–small cell lung cancer: a multivariate analysis. Can J Surg 2001;44:180-188.[Medline]
-
Harrell FE. Design. S-Plus functions for biostatistical/epidemiologic modeling, testing, estimation, validation, graphics, prediction, and typesetting by storing enhanced model design attributes in the fit. Available from http://lib.stat.cmu.edu or http://hesweb1.med.virginia.edu/s/Design.html.
-
S-Plus 2000, MathSoft, Inc., Seattle, WA.
-
Stone C.J., Koo C.Y. Additive splines in statistics. Proceedings of the Statistical Computing Section ASA. 1985:45-48.
-
Carles J., Rosell R., Ariza A., et al. Neuroendocrine differentiation as a prognostic factor in non–small cell lung cancer. Lung Cancer 1993;10:209-219.[Medline]
-
Lipford E.H., Eggleston J.C., Lillemoe K.D., Sears D.L., Moore G.W., Baker S. Prognosis factors in surgically resected limited-stage nonsmall cell carcinoma of the lung. Am J Surg Pathol 1984;8:357-365.[Medline]
-
Rivnay B., Gorelik E., Segal S., Shinitzky M. Plasma membrane microviscosity of Lewis lung carcinoma cells derived from local growth and pulmonary metastases. Invasion Metastasis 1981;1:99-110.[Medline]
-
Van Blitterswijk W.J., Hilkman H., Hengeveld T. Differences in membrane lipid composition and fluidity of transplanted GRSL lymphoma cells, depending on their site of growth in the mouse. Biochim Biophys Acta 1984;778:521-529.[Medline]
-
entjurc M., Sok M., Sera G. Plasma membrane fluidity alterations in cancerous tissues. Radiol Oncol 1998;32:109-117.
Top
Abstract
Introduction
