GYNECOLOGIC
ONCOLOGY
36, 23-29 (@!%)
Ultrastructural and lmmunohistochemical Study of Infiltration in Microinvasive Carcinoma of the Uterine Cervix RYUICHI KUDO,’ Department
of Obstetrics
and Gynecology,
TOMOKO Sapporo
SATO, AND HIDEMITSU
Medical
College,
South
MIZUUCHI
I, West 16, Chuo-ku,
Sapporo
060, Japan
Received November 4, 1988
Carcinomain situ and microinvasivecancerof the cervix were comparedby transmissionelectron microscopyto examineultrastructural featuresof the locally infiltrating lesionof mircoinvasive cancer.Many pseudopod-likecytoplasmicprotrusionsof the cancer cellsand abundant microfilamentsparallel to the direction of the protrusion were seen.Concomitant with the disapperanceof part of the basallamina, many vesicles70-90 nm in diameter were observed,suggestinga role for thesevesiclesin cancerinfiltration. With the immunoperoxidasemethod, the distribution of fibronectin around the invasive lesionalsowas examined.Fibronectin is a componentof extracellular matricesand presumably, in view of its action on cell adhesion,is a resistantfactor againstcancercell infiltration. Fibronectin decreasedin the transitional area betweenthe cancer nest and the stroma during the stageof microinvasion. 0 1990 Academic Press, Inc. INTRODUCTION
Investigation into the infiltration processes of cancer cells may be the first step toward understanding the biological behavior of the cancer. In general, however, it is difficult to observe the early infiltration process in human cancer tissue. One exception is uterine cervical cancer. In the present study, we investigated changes that may occur at the early stage of cancer cell infiltration by comparing microinvasive cancer of the uterine cervix with carcinoma in situ (CIS) or invasive cancer. The ultrastructural features of cancer cells at the front edge of the infiltration of microinvasive cancer, particularly of cancer cells adjacent to the stroma, were first examined by transmission electron microscopy (TEM). Localization of microfilaments (Mf) in microinvasive regions was detected by immunohistochemical staining with antiactin antibody. Fibronectin (FN), which is a component of extracellular matrices, is known to have ’ To whom requests for reprints should be addressed.
a cell-adhering action and to markedly decrease in amount with the transformation of cultured cells [l]. Fibronectin is also thought to be related to cancer infiltration or metastasis in vivo. The distribution pattern of FN, particularly the decrease in FN in the stroma around the margin of the cancer infiltration, was also examined. MATERIALS
AND METHODS
Ultrastructural study. The tissue specimens were obtained surgically at our department from 12 patients with microinvasive cancer, 25 patients with CIS, and 20 patients with invasive cancer. The histological type of all these specimens was squamous cell carcinoma. The diagnosis of microinvasive carcinoma was based on the definition proposed by the Japanese Society of Obstetrics and Gynecology and the Japanese Society of Pathology [2]. In this proposal, microinvasive carcinoma is defined as squamous cell carcinoma with histologically confirmed stromal invasion to a depth of 3 mm or less, as measured from the basement membrane of the overlying surface layer, and in which vascular involvement or confluent invasion is not demonstrated. The lesion that was judged to be the most serious by preoperative colposcopy was isolated with a surgical knife immediately after surgical resection of the cancer. The stromal portion was removed from the isolated tissue so as not to disturb the original cytoarchitecture. The tissue was sectioned to approximately 10x 5 x0.5 mm and fixed in 2.5% glutaraldehyde at room temperature. The buffer solution used for the fixation was the microtubule (Mt) condensation buffer reported by Luftig et al. [3], consisting of 1 mM guanosine triphosphate (GT), 1 mM MGSO, 2 mM ethylene glycol bis(@aminoethyl ether)-N,N-tetraacetic acid (EGTA), and 100 mM piperazine-N,iV’-bis(2-ethanesulfonic acid) (Pipes), pH 6.9. The tissue section must be vertical to the epithelial layer; therefore, the tissue sample was embedded in a flat
23 00?%8258/90$1.50 Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
KUDO,
SATO,
embedding plate which was attached to a cylindrical block. Toluidine blue-stained sections for light microscopic observation were prepared. After the microinvasive region was confirmed, the tissue was carefully trimmed, ultrasectioned, double-stained with lead and uranium, and observed in the electron microscope (JEM1200EX). Localization of Mf in the microinvasive region. The formalin-fixed paraffin sections from the same patient used for TEM observation were deparaffinized and treated with 3% hydrogen peroxide-methanol to inhibit the activity of endogeneous peroxidase. The sections were treated with normal goat serum to suppress nonspecific reactivity, and finally stained by the avidin-biotin-peroxidase complex method (ABC method) with rabbit antiactin antibody (1: 100 dilution, Miles Scientific) and the Vectastain kit (Vector) in a conventional manner. Control samples were treated with nonimmunized rabbit serum instead of the antiactin antibody. Localization of FN. Forty-five patients with microinvasive cancer, 21 patients with CIS, and 30 patients with invasive cancer who underwent surgical operations at our department from 1980 to 1987 were selected. The surgical specimens obtained from these patients by either surgical resection or biopsy were used. All of these tissues were diagnosed as squamous cell carcinoma. After removal of paraffin, formalin-fixed tissue specimens were treated with 3% hydrogen peroxide and normal goat serum, then stained in a conventional manner by the ABC method with rabbit anti-human FN antibody (I:100 dilution, Biomedical Technologies) and the Vectastain kit. Control specimens were treated with nonimmunized rabbit serum instead of the anti-FN antibody.
AND
MIZUUCHI
portion with the basal lamina, even in the microinvasive region. No specific difference was noted in intracellular organelles between microinvasive cancer and CIS (Fig. 5). There was a tendency for desmosomes to decrease in number in cells in contact with the stroma at the infiltrating edge of the microinvasion, but no difference was observed in other cells from CIS. Distribution of Mf in the microinvasive region. In microinvasive cancer, actin was found to have intensely accumulated in the infiltrating edge (Fig. 6), suggesting the presence of many Mf. In the case of CIS, including the superficial layer, the entire cancer region was homogeneously and mildly stained, and no particular localization was noted. Distribution of FN. Both the basal membrane and the adjacent stromal region were well stained for FN in all specimens of normal squamous epithelium, dysplasia, and CIS (Fig. 7). However, the invasive portion and adjacent noninvasive CIS portion of the microinvasive cancer showed very different affinities to staining for FN. The CIS portion continuously and intensely stained from the basal membrane to the stroma, whereas the invasive portion always showed a decrease in staining in certain regions between the edge of the infiltration and the stroma (Figs. 8, 9). In the invasive cancer (Fig. IO), a reduction in staining at the boundary region between the cancer nest and stroma was more extensive than in microinvasive cancer. In particular, this reduction was most prominent in the case with scirrhous infiltration. A similar reduction in staining in the stroma was observed also in the infiltrative nests on the vaginal wall and paracervical connective tissue, in addition to the primary cancer nests.
RESULTS Ultrastructural findings. For the microinvasive cancers, in the cytoplasm of many cancer cells at the infiltrating edge and adjacent to the stroma, pseudopod-like protrusions were observed toward the stroma (Fig. 1). No basal lamina was observed in such regions. Many filaments 4 to 5 nm in diameter were present, running parallel with the cytoplasmic protrusions in the cytoplasm (Figs. 2, 3). Based on their diameters, these filaments were thought to be Mf (Fig. 4). In some cases, filaments with a diameter of about 25 nm were also observed parallel to the cytoplasmic protrusions, and were thought to represent Mt (Fig. 3). In the basal lamina-free portion of the cytoplasm of a cancer cell, a number of very fine vesicles with a diameter of 70 to 90 nm were present (Figs. l-3). Some were open and in contact with the stroma (Figs. 2,3). Vesicles were also abundant at the infiltrating edge in invasive cancer tissues. but were rarely observed in the
DISCUSSION Microjilaments and microtubules in the microinvasive region. The microfilament, one of the constituents of the cell skeleton, is thought to be involved with morphological maintenance and movements of the cell [41. Moreover, in the case of cultured cells, the stress fiber, which is a bundle of microfilaments, disappears at the time of transformation; therefore, the microfilament is said to be related to the carcinogenesis of cells [A. In the present study, cytoplasmic protrusions from cancer cells were observed simultaneously with a decrease in desmosomes in the early infiltration portion of the cancer, and many microfilaments were found to be oriented parallel to the direction of this cytoplasmic protrusion, namely, that of infiltration. Thus, it was inferred that cancer cells infiltrate by the amoebalike movements of the cytoplasm. In addition, these microfilaments were
MICROINVASIVE
CANCER OF THE CERVIX
FIG. 1. Transmission electron micrograph of microinvasive cancer in the uterine cervix. Many pseudopod-like the cancer cell are seen together with many vesicles (arrow). ST, stroma. 9400x.
25
cytoplasmic protrusions of
FIG. 2. Transmission electron micrograph of microinvasive cancer. High magnification of the area of cytoplasmic protrusion. Abundant microfilaments (mf) are seen parallel to the direction of the protrusion. In accordance with the disappearance of part of the basal lamina, vesicles (arrow) 70-90 nm in diameter are also seen. 66,000~.
26
KUDO, SATO, AND MIZUUCHI
FIG. 3. Transmission electron micrograph of microinvasive cancer. High magnification of the area where many vesicles(V) are seen. ML microtubule; Mf, microfilaments. 72,000 x .
considered to be actin filaments by immunohistochemical staining. Cancer infiltration and vesicles in the region with absent basal lamina. The absence of vesicles with a diameter of 70-90 nm in CIS and in the region of the basal lamina, and their presence in the region without basal lamina in microinvasive cancer and in the edge portion of infiltration in invasive cancer, suggest that vesicles are closely related to infiltration. It is unclear from TEM findings alone, however, whether these vesicles are concerned with endocytosis or exocytosis. It may be reasonable to consider these vesicles as pinocytotic vesicles
FIG. 4. Transmission electron micrograph of microinvasive cer. High magnification of microfilaments (mf). 180,000 x .
can-
resulting from endocytosis on the basis of their size. On the other hand, the relationship of the vesicles to exocytosis may be indicated by the finding that transformed cultured cells produce the enzyme capable of decomposing type IV collagen [6], another important constituent of the basal lamina. Thus, the possibility that the vesicle contains various decomposition products of extracellular substrates cannot be ignored. Identification of the substances contained in the vesicle and immunoelectron microscopic studies on the function of the vesicle are to be carried out in future investigations. Distribution of FN. It has been shown that FN is markedly decreased in the fibroblast transformed by virus [l] or a carcinogenic substance and in explant tissue cultures from malignant tumor tissues. As possible causes of this FN decrease, the reduction in FN synthesis by the cell itself and decreased ability to bind to the cell surface have been proposed. In addition, enhanced destruction of FN has also been suggested [7,8]. That the cultured cell transformed by virus produces an FN-decomposing enzyme by itself has been reported [91. Moreover, FN is thought to be closely related to the infiltration and metastasis of cancer in vivo, and it has been reported to decrease in the region from the basal lamina to the surrounding stroma in infiltrative cancer tissues. However, no such investigations have been carried out on microinvasive cancer. Because the FN stain-
MICROINVASIVE
21
CANCER OF THE CERVIX
FIG. 5. Transmission electron micrograph of carcinoma in situ. Basal lamina(BL)
is present continuously.
22,000X.
ing of the region from the basal lamina to the stroma is preserved in CIS, whereas it is lost in invasive cancer, Labat-Robert et al. [lo] suggested consideration of an atypical lesion or differences in the distribution pattern of FN, but no mention was made regarding the FN pattern in the case of early invasion. In the present study, microinvasive cancer showed a decrease in FN staining in a narrow region as compared with invasive cancer. Thus, it was confirmed that FN
decreased during early infiltration. FN has also been considered to be useful for differential diagnosis between CIS and microinvasive cancer, because its presence or absence indicates the presence or absence of infiltration. In conclusion, there are differences between microinvasive cancer and CIS ultrastructurally and in the distribution of extracellular matrices, such as fibronectin. These findings may indicate a kind of biological behavior in this cancer.
FIG. 6. Immunoperoxidase staining for action in microinvasive cancer. Positive staining is intense in the microinvasive lesion (arrow). 140x.
FIG. 7. Immunoperoxidase staining for fibronectin in carcinoma in situ. There is no decrease in fibronectin in the transitional area between the cancer nest and the stroma. 170 x .
28
KUDO, SATO, AND MIZUUCHI
FIG. 8. Immunoperoxidase staining for fibronectin in microinvasive cancer. There is a decrease in fibronectin in the transitional area beta the cancer nest and the stroma (ST). 170x.
FIG. 9. Immunoperoxidase sive cancer. 170 x .
staining for fibronectin in microinva-
FIG. 10. Immunoperoxidase staining for invasive cancer. The, area of decreased flbronectin around the invasive lesion is broader than that of microinvasion. Ca, invasive cancer; St, stroma. 85 X .
MICROINVASIVE
CANCER OF THE CERVIX
C. M., and Shafie, S. Metastatic potential correlates with enzymatic degradation of basement membrane collagen, Nature (London) 284, 67-68 (1980).
REFERENCES 1.
2.
3.
4.
5. 6.
Chen, L. B., Gallimore, P. H., and McDougall, J. K. Correlation between tumor induction sensitive protein on the cell surface, Proc. Natl. Acad. Sci. USA 13, 3570-3574 (1976). The Joint Study Committee on Stage Ia Cancer in the Uterine Cervix. A new proposal regarding criteria for stage IA cancer in the uterine cervix, Gynecol. Oncol. 8, 353-369 (1979). Luftig, R. B., McMillan, P. N., Weatherbee, J. A., and Weihing, R. R. Increased visualization of microtubules by improved fixation procedure, J. Histochem. Cytochem. 25, 175-187 (1977). Rubino, S., Fighett, M., Unger, F., and Cappucinelli, P. Location of actin, myosin and microtubular structures during directed locomotion of Dichtyosteliutn amebae, J. Cell Biol. 98, 382-390 (1984). Ben-Zeev, A. The cytoskeleton in cancer cells, Biochim. Biophys. Acta 780, 197-212 (1985). Liotta, L. A., Tryggvason, K., Garbisa, S., Hart, I., Foltz,
29
Hynes, R. O., and Wyke, J. A. Alterations in surface proteins in chicken cells transformed by temperature-sensitive mutants of Rous sarcoma virus, Virology 64, 492-504 (1975). Olden, K., and Yamada, K. M. Mechanism of the decrease in the major cell surface protein of chick embryo fibroblasts after transformation, Cell 11, 957-969 (1977). Miyazaki, K., J., and Horio, sarcoma virus with secretion fibronectin, J. 10.
Ashida, Y., Kihira, Y., Mashima, K., Yamashita, T. Transformation of rat liver cell line by Rous causes loss of cell surface fibronectin accompanied of metallo-proteinase that preferentially digests the Biochem. 102, 569-582 (1987).
Labat-Robert, J., Birembaut, P., Robert, L., and Adnet, J. J. Modification of the fibronectin distribution pattern in solid human tumors, Diagn. Histopathol. 4, 299-306 (1981).
ONCOLOGY
36, 23-29 (@!%)
Ultrastructural and lmmunohistochemical Study of Infiltration in Microinvasive Carcinoma of the Uterine Cervix RYUICHI KUDO,’ Department
of Obstetrics
and Gynecology,
TOMOKO Sapporo
SATO, AND HIDEMITSU
Medical
College,
South
MIZUUCHI
I, West 16, Chuo-ku,
Sapporo
060, Japan
Received November 4, 1988
Carcinomain situ and microinvasivecancerof the cervix were comparedby transmissionelectron microscopyto examineultrastructural featuresof the locally infiltrating lesionof mircoinvasive cancer.Many pseudopod-likecytoplasmicprotrusionsof the cancer cellsand abundant microfilamentsparallel to the direction of the protrusion were seen.Concomitant with the disapperanceof part of the basallamina, many vesicles70-90 nm in diameter were observed,suggestinga role for thesevesiclesin cancerinfiltration. With the immunoperoxidasemethod, the distribution of fibronectin around the invasive lesionalsowas examined.Fibronectin is a componentof extracellular matricesand presumably, in view of its action on cell adhesion,is a resistantfactor againstcancercell infiltration. Fibronectin decreasedin the transitional area betweenthe cancer nest and the stroma during the stageof microinvasion. 0 1990 Academic Press, Inc. INTRODUCTION
Investigation into the infiltration processes of cancer cells may be the first step toward understanding the biological behavior of the cancer. In general, however, it is difficult to observe the early infiltration process in human cancer tissue. One exception is uterine cervical cancer. In the present study, we investigated changes that may occur at the early stage of cancer cell infiltration by comparing microinvasive cancer of the uterine cervix with carcinoma in situ (CIS) or invasive cancer. The ultrastructural features of cancer cells at the front edge of the infiltration of microinvasive cancer, particularly of cancer cells adjacent to the stroma, were first examined by transmission electron microscopy (TEM). Localization of microfilaments (Mf) in microinvasive regions was detected by immunohistochemical staining with antiactin antibody. Fibronectin (FN), which is a component of extracellular matrices, is known to have ’ To whom requests for reprints should be addressed.
a cell-adhering action and to markedly decrease in amount with the transformation of cultured cells [l]. Fibronectin is also thought to be related to cancer infiltration or metastasis in vivo. The distribution pattern of FN, particularly the decrease in FN in the stroma around the margin of the cancer infiltration, was also examined. MATERIALS
AND METHODS
Ultrastructural study. The tissue specimens were obtained surgically at our department from 12 patients with microinvasive cancer, 25 patients with CIS, and 20 patients with invasive cancer. The histological type of all these specimens was squamous cell carcinoma. The diagnosis of microinvasive carcinoma was based on the definition proposed by the Japanese Society of Obstetrics and Gynecology and the Japanese Society of Pathology [2]. In this proposal, microinvasive carcinoma is defined as squamous cell carcinoma with histologically confirmed stromal invasion to a depth of 3 mm or less, as measured from the basement membrane of the overlying surface layer, and in which vascular involvement or confluent invasion is not demonstrated. The lesion that was judged to be the most serious by preoperative colposcopy was isolated with a surgical knife immediately after surgical resection of the cancer. The stromal portion was removed from the isolated tissue so as not to disturb the original cytoarchitecture. The tissue was sectioned to approximately 10x 5 x0.5 mm and fixed in 2.5% glutaraldehyde at room temperature. The buffer solution used for the fixation was the microtubule (Mt) condensation buffer reported by Luftig et al. [3], consisting of 1 mM guanosine triphosphate (GT), 1 mM MGSO, 2 mM ethylene glycol bis(@aminoethyl ether)-N,N-tetraacetic acid (EGTA), and 100 mM piperazine-N,iV’-bis(2-ethanesulfonic acid) (Pipes), pH 6.9. The tissue section must be vertical to the epithelial layer; therefore, the tissue sample was embedded in a flat
23 00?%8258/90$1.50 Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
KUDO,
SATO,
embedding plate which was attached to a cylindrical block. Toluidine blue-stained sections for light microscopic observation were prepared. After the microinvasive region was confirmed, the tissue was carefully trimmed, ultrasectioned, double-stained with lead and uranium, and observed in the electron microscope (JEM1200EX). Localization of Mf in the microinvasive region. The formalin-fixed paraffin sections from the same patient used for TEM observation were deparaffinized and treated with 3% hydrogen peroxide-methanol to inhibit the activity of endogeneous peroxidase. The sections were treated with normal goat serum to suppress nonspecific reactivity, and finally stained by the avidin-biotin-peroxidase complex method (ABC method) with rabbit antiactin antibody (1: 100 dilution, Miles Scientific) and the Vectastain kit (Vector) in a conventional manner. Control samples were treated with nonimmunized rabbit serum instead of the antiactin antibody. Localization of FN. Forty-five patients with microinvasive cancer, 21 patients with CIS, and 30 patients with invasive cancer who underwent surgical operations at our department from 1980 to 1987 were selected. The surgical specimens obtained from these patients by either surgical resection or biopsy were used. All of these tissues were diagnosed as squamous cell carcinoma. After removal of paraffin, formalin-fixed tissue specimens were treated with 3% hydrogen peroxide and normal goat serum, then stained in a conventional manner by the ABC method with rabbit anti-human FN antibody (I:100 dilution, Biomedical Technologies) and the Vectastain kit. Control specimens were treated with nonimmunized rabbit serum instead of the anti-FN antibody.
AND
MIZUUCHI
portion with the basal lamina, even in the microinvasive region. No specific difference was noted in intracellular organelles between microinvasive cancer and CIS (Fig. 5). There was a tendency for desmosomes to decrease in number in cells in contact with the stroma at the infiltrating edge of the microinvasion, but no difference was observed in other cells from CIS. Distribution of Mf in the microinvasive region. In microinvasive cancer, actin was found to have intensely accumulated in the infiltrating edge (Fig. 6), suggesting the presence of many Mf. In the case of CIS, including the superficial layer, the entire cancer region was homogeneously and mildly stained, and no particular localization was noted. Distribution of FN. Both the basal membrane and the adjacent stromal region were well stained for FN in all specimens of normal squamous epithelium, dysplasia, and CIS (Fig. 7). However, the invasive portion and adjacent noninvasive CIS portion of the microinvasive cancer showed very different affinities to staining for FN. The CIS portion continuously and intensely stained from the basal membrane to the stroma, whereas the invasive portion always showed a decrease in staining in certain regions between the edge of the infiltration and the stroma (Figs. 8, 9). In the invasive cancer (Fig. IO), a reduction in staining at the boundary region between the cancer nest and stroma was more extensive than in microinvasive cancer. In particular, this reduction was most prominent in the case with scirrhous infiltration. A similar reduction in staining in the stroma was observed also in the infiltrative nests on the vaginal wall and paracervical connective tissue, in addition to the primary cancer nests.
RESULTS Ultrastructural findings. For the microinvasive cancers, in the cytoplasm of many cancer cells at the infiltrating edge and adjacent to the stroma, pseudopod-like protrusions were observed toward the stroma (Fig. 1). No basal lamina was observed in such regions. Many filaments 4 to 5 nm in diameter were present, running parallel with the cytoplasmic protrusions in the cytoplasm (Figs. 2, 3). Based on their diameters, these filaments were thought to be Mf (Fig. 4). In some cases, filaments with a diameter of about 25 nm were also observed parallel to the cytoplasmic protrusions, and were thought to represent Mt (Fig. 3). In the basal lamina-free portion of the cytoplasm of a cancer cell, a number of very fine vesicles with a diameter of 70 to 90 nm were present (Figs. l-3). Some were open and in contact with the stroma (Figs. 2,3). Vesicles were also abundant at the infiltrating edge in invasive cancer tissues. but were rarely observed in the
DISCUSSION Microjilaments and microtubules in the microinvasive region. The microfilament, one of the constituents of the cell skeleton, is thought to be involved with morphological maintenance and movements of the cell [41. Moreover, in the case of cultured cells, the stress fiber, which is a bundle of microfilaments, disappears at the time of transformation; therefore, the microfilament is said to be related to the carcinogenesis of cells [A. In the present study, cytoplasmic protrusions from cancer cells were observed simultaneously with a decrease in desmosomes in the early infiltration portion of the cancer, and many microfilaments were found to be oriented parallel to the direction of this cytoplasmic protrusion, namely, that of infiltration. Thus, it was inferred that cancer cells infiltrate by the amoebalike movements of the cytoplasm. In addition, these microfilaments were
MICROINVASIVE
CANCER OF THE CERVIX
FIG. 1. Transmission electron micrograph of microinvasive cancer in the uterine cervix. Many pseudopod-like the cancer cell are seen together with many vesicles (arrow). ST, stroma. 9400x.
25
cytoplasmic protrusions of
FIG. 2. Transmission electron micrograph of microinvasive cancer. High magnification of the area of cytoplasmic protrusion. Abundant microfilaments (mf) are seen parallel to the direction of the protrusion. In accordance with the disappearance of part of the basal lamina, vesicles (arrow) 70-90 nm in diameter are also seen. 66,000~.
26
KUDO, SATO, AND MIZUUCHI
FIG. 3. Transmission electron micrograph of microinvasive cancer. High magnification of the area where many vesicles(V) are seen. ML microtubule; Mf, microfilaments. 72,000 x .
considered to be actin filaments by immunohistochemical staining. Cancer infiltration and vesicles in the region with absent basal lamina. The absence of vesicles with a diameter of 70-90 nm in CIS and in the region of the basal lamina, and their presence in the region without basal lamina in microinvasive cancer and in the edge portion of infiltration in invasive cancer, suggest that vesicles are closely related to infiltration. It is unclear from TEM findings alone, however, whether these vesicles are concerned with endocytosis or exocytosis. It may be reasonable to consider these vesicles as pinocytotic vesicles
FIG. 4. Transmission electron micrograph of microinvasive cer. High magnification of microfilaments (mf). 180,000 x .
can-
resulting from endocytosis on the basis of their size. On the other hand, the relationship of the vesicles to exocytosis may be indicated by the finding that transformed cultured cells produce the enzyme capable of decomposing type IV collagen [6], another important constituent of the basal lamina. Thus, the possibility that the vesicle contains various decomposition products of extracellular substrates cannot be ignored. Identification of the substances contained in the vesicle and immunoelectron microscopic studies on the function of the vesicle are to be carried out in future investigations. Distribution of FN. It has been shown that FN is markedly decreased in the fibroblast transformed by virus [l] or a carcinogenic substance and in explant tissue cultures from malignant tumor tissues. As possible causes of this FN decrease, the reduction in FN synthesis by the cell itself and decreased ability to bind to the cell surface have been proposed. In addition, enhanced destruction of FN has also been suggested [7,8]. That the cultured cell transformed by virus produces an FN-decomposing enzyme by itself has been reported [91. Moreover, FN is thought to be closely related to the infiltration and metastasis of cancer in vivo, and it has been reported to decrease in the region from the basal lamina to the surrounding stroma in infiltrative cancer tissues. However, no such investigations have been carried out on microinvasive cancer. Because the FN stain-
MICROINVASIVE
21
CANCER OF THE CERVIX
FIG. 5. Transmission electron micrograph of carcinoma in situ. Basal lamina(BL)
is present continuously.
22,000X.
ing of the region from the basal lamina to the stroma is preserved in CIS, whereas it is lost in invasive cancer, Labat-Robert et al. [lo] suggested consideration of an atypical lesion or differences in the distribution pattern of FN, but no mention was made regarding the FN pattern in the case of early invasion. In the present study, microinvasive cancer showed a decrease in FN staining in a narrow region as compared with invasive cancer. Thus, it was confirmed that FN
decreased during early infiltration. FN has also been considered to be useful for differential diagnosis between CIS and microinvasive cancer, because its presence or absence indicates the presence or absence of infiltration. In conclusion, there are differences between microinvasive cancer and CIS ultrastructurally and in the distribution of extracellular matrices, such as fibronectin. These findings may indicate a kind of biological behavior in this cancer.
FIG. 6. Immunoperoxidase staining for action in microinvasive cancer. Positive staining is intense in the microinvasive lesion (arrow). 140x.
FIG. 7. Immunoperoxidase staining for fibronectin in carcinoma in situ. There is no decrease in fibronectin in the transitional area between the cancer nest and the stroma. 170 x .
28
KUDO, SATO, AND MIZUUCHI
FIG. 8. Immunoperoxidase staining for fibronectin in microinvasive cancer. There is a decrease in fibronectin in the transitional area beta the cancer nest and the stroma (ST). 170x.
FIG. 9. Immunoperoxidase sive cancer. 170 x .
staining for fibronectin in microinva-
FIG. 10. Immunoperoxidase staining for invasive cancer. The, area of decreased flbronectin around the invasive lesion is broader than that of microinvasion. Ca, invasive cancer; St, stroma. 85 X .
MICROINVASIVE
CANCER OF THE CERVIX
C. M., and Shafie, S. Metastatic potential correlates with enzymatic degradation of basement membrane collagen, Nature (London) 284, 67-68 (1980).
REFERENCES 1.
2.
3.
4.
5. 6.
Chen, L. B., Gallimore, P. H., and McDougall, J. K. Correlation between tumor induction sensitive protein on the cell surface, Proc. Natl. Acad. Sci. USA 13, 3570-3574 (1976). The Joint Study Committee on Stage Ia Cancer in the Uterine Cervix. A new proposal regarding criteria for stage IA cancer in the uterine cervix, Gynecol. Oncol. 8, 353-369 (1979). Luftig, R. B., McMillan, P. N., Weatherbee, J. A., and Weihing, R. R. Increased visualization of microtubules by improved fixation procedure, J. Histochem. Cytochem. 25, 175-187 (1977). Rubino, S., Fighett, M., Unger, F., and Cappucinelli, P. Location of actin, myosin and microtubular structures during directed locomotion of Dichtyosteliutn amebae, J. Cell Biol. 98, 382-390 (1984). Ben-Zeev, A. The cytoskeleton in cancer cells, Biochim. Biophys. Acta 780, 197-212 (1985). Liotta, L. A., Tryggvason, K., Garbisa, S., Hart, I., Foltz,
29
Hynes, R. O., and Wyke, J. A. Alterations in surface proteins in chicken cells transformed by temperature-sensitive mutants of Rous sarcoma virus, Virology 64, 492-504 (1975). Olden, K., and Yamada, K. M. Mechanism of the decrease in the major cell surface protein of chick embryo fibroblasts after transformation, Cell 11, 957-969 (1977). Miyazaki, K., J., and Horio, sarcoma virus with secretion fibronectin, J. 10.
Ashida, Y., Kihira, Y., Mashima, K., Yamashita, T. Transformation of rat liver cell line by Rous causes loss of cell surface fibronectin accompanied of metallo-proteinase that preferentially digests the Biochem. 102, 569-582 (1987).
Labat-Robert, J., Birembaut, P., Robert, L., and Adnet, J. J. Modification of the fibronectin distribution pattern in solid human tumors, Diagn. Histopathol. 4, 299-306 (1981).
