15, pp. 1353-1359,
Copyright0 1998Elscvier.Science Printed
in the USA.
HLA DR AND AB SURFACE ANTIGENS CORRELATE (SURFACE AREA)
WITH CELL SHAPE
Norman Ende’ 8~ Arthur B. Ritter” *Department of Pathology and Laboratory Medicine, and #Department of Pharmacology Physiology, UMD-New Jersey Medical School, Newark, NJ 07103, USA
(Received in linal form July 28, 1998) Summary In order to help explain some of the various phenomena associated with both benign and malignant cells, this study was undertaken to determine if changes in the shape of the cell could alter the recognition of the cell. Non-transformed Human cells, HEL 299, were evaluated for their shape and surface antigens. A direct statistical correlation was found between the two surface antigens HLA AB and DR and the cell shape (surface area). The possible significance of this phenomena in non transformed human cells to neoplastic proliferation is suggested. Key Wordvt
Background In 1979 we noted that in a study on cytotoxic kidney antibodies that human kidney cells directly obtained from renal tissues did not show the typical cytotoxicity effect or absorption effect when exposed to known anti-kidney serum (1). However, suspension of cells obtained following the first passage monolayer did produce cells that responded characteristically when exposed to anti-kidney serum both as related to cytotoxicity and absorption. Following these findings, it occurred to us that normal cells could change their surface antigens by merely changing shape. If normal cells in the process of routine repair and changing surface areas can alter their surface antigens, it is possible that during the process the host may or may not develop recognition of a proliferating clone of cells and potentially allow cells to escape from recognition. If this did occur, self-recognition could tentatively fail in normal cells but perhaps to a much greater frequency in neoplastic cells. Since variation in shape (surface area) is one of the key findings for the pathological diagnosis of a malignant neoplastic process, a similar phenomenon of shape affecting surface antigens could occur with malignant cells, but to a greater degree and make self recognition more difficult. Malignant cells have a variety of sizes and shapes, expose a complexity of surface antigens (2) and show diversity between different individuals (3). The following study was undertaken to determine if by merely altering the shape (surface area) of non-transformed cells, the number of surface antigens could be affected.
Vol. 63, No. 15, 1998
b 8 E z
Fig. 1. Fluorescent Antibody Analysis. a. The Video Image Of The Surface Antigens Of An Individual Cell Labeled With Fluorescence. b. Histogram Of The Individual Cell With Representation Of Background, Cell Body And Antibody Gray Colors On Cell Surface.
Vol. 53, No. 15, 1998
Cell Size (pixels)
. . . 3oooo
Cell Size (pixels)
Fig. 2. a. Antibody fluorescence VS Cell Surface Area For HLA-DR-Tagged Cells. b. Antibody Fluorescence VS Cell Surface Area For HLA-ABTagged Cells.
Methods Cell Type: Human embryonic lung HEL 299 was obtained from American Type Culture Collections. This is a non-transformed, adherent cell line and was maintained in Minimal Essential Media (Gibco Co.) with Earls salts; supplemented with 10% fetal bovine serum, 1% I,-glutamine. 1% penicillinistreptomycin/fungizone solution. Poly (HEMA) muting of.cover slides: To alter the shape of cells the cells were grown on cover slips previously coated with poly ((HEMA)2-hydroxyethylmethacrylate supplied by the Aldrich Chemical Company, Inc. The technique utilized for varying cell shape was similar to that described by Folkman (4) wherein the thickness of the cell varied with the concentration of the poly (HEMA). Dilutions of the poly (HEMA) least one hour. were incubated
from IO-* to 10m4. Complete
Cells were added
at a density
media were utilized
for 18 hours in a 37O C humidified
3 x 105 cells per well.
for at Trays
Fluorescent Stuins Monoclonal antibodies prepared in the mouse were utilized in the direct technique staining of the HLA antigens, HLA, AB and DR (accurate antibodies. Westburg, NY & U.S. Bioproducts, Inc., Queens Village, NY). The biotinylated antibody used was anti-mouse in the indirect method. The cover slips with attached cells were removed from the multi well plates. 150 ul of biotinylated monoclonai antibody were added to the cover slip and incubated on ice for 1.5 minutes. The bridge antibody was then removed and cover slips washed 3 times with cold Debco modified Eagle’s media. The cover slips, cells down, were placed on a slide with 10 ul of buffered glycerol. were then examined with a compound microscope with an attached T.V. camera.
Visualization The cells used in this study were human lung embryonic cells which were incubated with fluorescently - labeled (fluorescein) antibodies. There were two tagged antigens, AB and DR. All cells examined had undergone similar periods of incubation. A few drops of suspended cells were placed on a microscope slide and covered by a glass cover slip. The cells were epiilluminated through the microscope objective with incident light of 488 nm (UV) wavelength (13). The fluorachrome attached to the antibodies fluoresces at a peak wavelength of 505 nm (green). A dichrome mirrror and optical filter pass light at the fluorescent wavelength (505 nm) from the objective to a low light, silicon intensifier target (SIT) television camera (Cohu) on which the gain and pedestal controls were manually set ( the automatic gain control was turned off). A micrometer scale was also recorded in both vertical and horrizontal positions to calibrate absolute pixel size and area. Video images of the cells were recorded on 3/4 inch video tape (Sony - Umatic). The tape was played back frame by frame for analysis of surface antigenic markers as measured by fluorescent intensity on the cell surface of individual cells (11). 44 AB cells and 39 DR cells were analysed.
Vol. 63, No. 15, 1998
Surface Antigens Correlate
Processing of the data Only individual cells were selected from each frame and digitized into a 5 12x512~8 bit grayscale image using a video image digitizer (Quantex). The digitized data files were converted to Application Visualization System (AVS) field data format (Advanced Visual Systems, Inc.) using a small program written in C. The AVS software runs under UNIX on a Silicon Graphics workstation ( Indigo) which has high resolution graphics display capability. This allowed us to better identify the graycolor intensity ranges which represent the cell body, the fluorescent antibody and the non-cell background. Area Calculations The minimum and maximum threshold values of graylevel intensity for cell body, antibody and background were determined for each cell using the histogram function of AVS. Figure 1 is a typical graylevel histogram of a cell showing the different features. The maximum and minimum threshold values of the cell body, antibody and background were passed to a small program written in C which counts the number of pixels that lie in each of the various graylevel ranges, multiplies them by the calibrated pixel area and calculates the individual areas. Accuracy of the software was tested with simulated grayscale data of known areas. Statistical Analysis For each cell type, fluorescent (antibody) area was plotted against surface area (shape) and a correlation coefficient, r, calculated (12). We tested the significance of the correlation coefficients at the 0.05 level using the Fisher Z transformation (12). Briefly, Z is calculated as a In transformation of r : Z=+ln k=
where r is the correlation coeffficient, p is the population correlation coefficient, Z is the Fisher Z transformation, z is the number of standard deviations away from the mean value that the transformed r value lies on the standard normal curve, pz is the transformed population correlation coefficient and or is the population standard deviation. We test the null hypothesis p = 0 (no correlation) against the alternative hypothesis p # 0 at cx = 0.05. Since pL = 0 for p = 0, reject the null hypothesis if z 2 1.96 or z < - 1.96, where z is defined above. Results The relationship between surface area and surface antigens is plotted in figure 2 for both DR and AB-labeled cells. For the DR cells, r = 0.65, Z = 0.78, o, = 0.17 (n = 39) and z = 4.65. Since z > 1.96, we can reject the null hypothesis at the 0.05 level and conclude that the correlation is statistically significant.
Vol. 63, No. 15, 19%
Surface Antigens Corrclale Ccl1 Shape
For the AB cells, r = 0.70, Z = 0.87, cr< = 0.16 (n = 44) and z = 5.57. Since z > 1.96, we can reject the null hypothesis at the 0.05 level and conclude that the correlation is statistically significant. Discussion
Erlich postulated approximately 90 years ago that an organism can eliminate foci of abnormal cells (5). If, potentially, the phenomenon of escape from immunity can occur with normal cells, then in a similar fashion genetic instability could result in continuous production of neoplastic cells (6) that escape detection. An organism’s immunity to control this process should have a quantitative limit. The Gl
of the cell cycle
as the point
appearance. This is also the phase of blebbed appearance of the cells (7). Cell shape has been found to be closely coupled to DNA synthesis in non-transformed cells (4). If the antigens on the surface can be continuously modified by simply altering the surface area of the cells, this could further the possibility of escape from immunological control. In the process of any tissue repair, the replacement cells must undergo alterations in size and shape to replace the defect produced by the injury. This normal repair process, therefore, on occasion may allow a window of opportunity for cells to escape recognition simply by the change of surface area and thereby the surface antigens. The cells selected for these studies were Human Embryonic Lung Cells (HEL 299) which are Once not considered transformed and would be expected to function as normal cells. transformed, the cells would be expected to show variation of size, shape and surface antigens and therefore transformed cells were not utilized in this study. All malignancies are believed to be derived from normally functioning cells which undergo transformation into uncontrolled growth. Therefore, the behavior and recognition of normally functioning cells during this transitional phase would be critical in the development of a malignant process. Many solid neoplasms have been associated with some form of chronic repetitive injury and repair on the cellular level. Examples of these are carcinoma of the lung, esophagus and liver. Neoplastic cells once evolved, whose pathological hallmark is changes in size and shape, would be particularly prone to present different surface antigens and thereby more readily continue to If a change in surface area can thereby change the cells’ escape recognition and detection. recognition signals, it may help explain the value of debulking of large tumors, the failure of antibody therapy, the significance of the tumor size in obtaining cures, the “sneaking through phenomenon” and other phenomenon associated with solid tumor malignancies (8,9,10). Acknowledgement
in part by the Abraham
S. Ende Research Foundation
and Norton Lilly International.
I wish to thank Ernest V. Orsi, Ph.D., Nicholas M. Ponzio, Ph.D., Allen McKenzie, Dona Cole and Rosanna Ricafort and others, not mentioned, for their assistance on this study over the past 20 years. We also wish to thank Amos Gona, Ph.D. and Elizabeth Raveche for reviewing the manuscript. References
N. ENDE, E.V. ORSI, N. Z. BATIJRAY, 543-548 (1979).
and T. L. BRITTENY,
Am J Clin Path 71
Vol. 63, No. 15, 199)X
2. 3. 4. 5. 6. 7. 8. 9.
10. 11. 12. 13.
C. TING, and R. B. HERBERMAN, Nature 257 801-802 (1975). P. HERSHEY, Aus N Z Med 7 526-536 ( 1977). J. FOLKMAN and A. MOSCONA, Nature 273 345-349 (1978). P. EHRLICH, The collectedpapers ofPaul Ehrlich, F. Himmelweit (Ed), p. 550, London Pergamon Press (1957). F.I. DREW, Progress in Human Pathology 10 5-14 (1979). K. H. BURK and B. DREWINKO, Cancer Research 36 3535-3538 (1976). R.T. SMITH, Advances in Pathology 4 4-85 (1976). K. TAMAKA , T. YOSHIOKA, C. BIERBERICH and G. JAY G. Am Rev Immuno16 359-80 (1955). E.F. WHEELOCK, and M. K. ROBINSON, Laboratory Inves. 48 120-139 (1985). M. GOLDMAN, Fluorescent Antibody Methods. Bionetics Research Lab., Inc. Falls Church VA. Academic Press, New York, (1968). J.E FREUND and G.A. SIMON, Prentice-Hall, Inc. Englewood Cliffs, NJ (1992). A. B. RITTER, W. BRAUN, A. STEIN, and W. DURAN, Computers in Biology and Medicine 15 361-374 (1985).