Breast Cancer Research and Treatment 22:31-38, 1992. © 1992 Kluwer Academic Publishers. Printed in the Netherlands.
Expression of IGF-I and IGF-II mRNA in breast tissue Soonmyoung Paik Department of Pathology, Lombardi Cancer Center, Georgetown University, Washington DC, USA
Key words: breast tumor cells, epithelial cells, IGF-I, IGF-II, in-situ hybridization, insulin-like growth factors, messenger RNA, stromal cells
Summary Expression of IGF-I and IGF-II was studied in human breast cancer tissues by in situ hybridization. IGF-I mRNA was detected only in stromal cells adjacent to normal breast epithelial cells. Stromal cells associated with the tumor cells did not contain IGF-I, nor did malignant or benign breast epithelial ceils. In contrast, IGF-II mRNA was found in both the malignant epithelial cells and their adjacent stromal cells. These data imply that stromal cells associated with breast epithelium may switch expression from IGF-I to IGF-II during breast cancer evolution. This appearance of IGF-II expression may identify cancerassociated stromal cells that have a fetal phenotype.
Introduction Although earlier in-vitro studies suggested that IGF-I might be an estrogen inducible autocrine growth factor for breast cancer cells, measurement of IGF-I mRNA levels in homogenized breast tissue specimens using RNase protection assay showed an unexpected result - - there was decreased expression of IGF-I mRNA in malignant tissues when compared to adjacent normal tissues [1]. This data, combined with the failure to demonstrate IGF-I expression in breast cancer cell lines, suggested that IGF-I was unlikely to be produced by malignant epithelial cells. Several alternative hypotheses were developed to explain the detection of IGF-I in breast cancer tissues: 1) IGF-I mRNA is expressed by normal roam-
mary epithelial cells but not by malignant cells; or 2) stromal cells in the normal lobule produce IGF-I while stromal cells associated with the cancer do not. In order to distinguish between these two possibilities, we have used in-situ hybridization to study the expression of IGF-I mRNA in paraffin-embedded sections.
Methods Probes
The IGF-I cDNA was kindly provided by Ken Gabby (Baylor College of Medicine, Houston, TX). The 540bp PstI-BamHI fragment of IGF-IA cDNA was subcloned into pGem4 (Promega) after
Addressfor offprints: SoonmyoungPaik, M.D., Room S-165-A,LombardiCancerCenter, 3800 ReservoirRd NW, Washington DC 20007, USA
32
S Paik
removal of 21bp of the 5' poly(G) tail. IGF-II cDNA was a kind gift from Graeme Bell (Howard Hughes Institute, Chicago, IL). The 854bp PstIPstI fragment from the coding region was cloned into pGem4. Labeled antisense RNA was transcribed using T7 polymerase according to instructions of the manufacturer. Labeled sense RNA was transcribed using SP6 polymerase. 35SUTP (NEN, Wilmington, DE) was triple concentrated by lyophilization before synthesis of the probe.
Hybridization protocol Previously described methods [2] were used except for modifications of the pretreatment step for paraffin-embedded sections. All pre-ribonuclease (RNase) washing steps were carried out using autoclaved polypropylene staining dishes (Miles Lab). Sections were deparaffinized and rehydrated through alcohol series. Basic proteins were removed by incubating 20 minutes in 0.2M HCI. After washing in distilled water for 5 minutes, sections were denatured by incubating 30 minutes at 70°C in 2xSSC. Sections were then digested with 0.5 /ag/ml proteinase K in 20mM Tris-HC1/2mM CaC12 for 3 minutes at room temperature. Sections were then incubated 30 minutes in 0.1M Tris-0.1M Glycine, washed twice for 3 minutes each in Dutbecco's PBS, and acetylated for 10 minutes in fresh acetic anhydride diluted 1:400 in 0.1M triethanolamine (pH 8.0). Sections were then washed in Dulbecco's PBS for 5 minutes and dehydrated through 30%, 60%, 80%, 95%, and 100% ethanol for 2 minutes each and air dried for 1 hour. Hybridization mix contained 50% formamide, 10% dextran sulfate, 2xSSC, 2mg/ml bovine serum albumin, lmg/ml salmon sperm DNA, lmg/ml yeast tRNA, 50raM DTT, and 5X10 7 cpm/ml of probe. Probe mix was prepared and kept in a 50°C water bath. Sections were covered with 100 to 200 pl of hybridization mix by careful spreading with a pipette tip with-
out touching the section. No coverslip was used to prevent tissue detachment. Hybridization time was 4 hours. Sections were transferred to 50% formamide/2xSSC at 52°C to wash off hybridization mix, washed in 50% formamide/2xSSC with 10raM DTT at 52°C for 5 minutes with continuous vertical washing, then washed for 20 minutes in a new batch of same solution, washed 4 times 1 minute in 2xSSC at room temperature, incubated in 20~g/ml RNase A prepared in 2xSSC for 30 minutes at 37°C, and rinsed twice 1 minute each in 2xSSC. Final wash was done in 50% formamide/2xSSC without DTT at 60°C for 30 minutes with continuous agitation. After washing 1 minute in 2xSSC, sections were dehydrated through 70%, 80%, and 95% ethanol for 1 minute each and air dried for 1 hour. Sections were then exposed to Hyperfilm-beta max (Amersham) for 3 days to check the completeness of washing, and if necessary rehydrated and washed again in 0. l xSSC for 30 minutes with continuous agitation and dehydrated. Sections were then exposed to NTB-2 emulsion for 3 weeks, developed for 2.5 minutes in Kodak developer, fixed for 3 minutes in Kodak fixer, washed for 5 minutes in distilled water, and air dried. Sections were then stained with hematoxylin and eosin, dehydrated, coverslipped, and examined under a microscope. After examination and photography, cover slips were removed and the sections were re-exposed for 3 weeks to Hyperfilm-beta max. Emulsion exposure and film exposure were then compared to rule out possible artifacts in emulsion exposure.
Optimization of" the assay conditions The assay was optimized using cell lines known to express only one of the IGF species: cells expressing only IGF-I mRNA (CHP-100, American Type Culture Collection, Rockville, MD), and cells which express only IGF-II mRNA (HepG2, ATCC). Time course of the hybridization reaction showed maximum signal to noise ratio
IGF mRNA expression in breast tissue
33
hybridization. Note strong black signal over CHP-100 cells and no signal over HepG2 cells. A case of tubular adenoma with known m R N A expression by RNase protection assay was fixed overnight in 10% formalin and embedded in paraffin. This section was used as a positive control for subsequent assays.
Figure 1. Example of film exposure for IGF-I in-situ hybridization. IGF-I expressing CHP100 cells show strong hybridization signal whereas IGF-II expressing control cells (HepG2) show no signal. after 4 hours at 52°C. Background signal increased exponentially after 4 hours hybridization without increase in specific hybridization signal. Washing conditions were optimized by exposing the sections to the beta-sensitive film (Hyperfilm beta max) before emulsion exposure. Figure 1 is an example of film exposure of positive and negative control cell lines for IGF-I in-situ
Results Expression of IGF-I mRNA in human breast cancer None of the sections from 10 breast cancer cases (without normal lobules in the section) showed significant IGF-I expression. However, in two cases in which a normal lobular component was associated with the breast cancer, expression of IGF-I mRNA was detected limited to stromal cells in the non-malignant area (Figure 2). The
Figure 2. Expression of IGF-I mRNA in infiltrating ductal carcinoma of the breast (100X). Note hybridization signal over stromal cells in non-malignant areas (N) whereas there are no signals over either malignant epithelial cells or stromal cells around them (T).
34
S Paik
Figure 3. A) Expression of IGF-I mRNA in normal lobule (100X). Hybridization signals are confined over a subset of stromal cells (S) in the Iobule. B) Positive cells are more abundant in edematous area (E), C) Expression of IGF-I mRNA in fibroadenoma (100X). All hybridization signals are confined to stromal component (S) and not epithelial cells (E).
IGF mRNA expression in breast tissue
35
Figure 4. Expressionof IGF-II mRNA in non-malignantbreast tissue, in this case with blunt duct adenosis (dark field photomicrograph, 100X magnification). Hybridizationsignals are over stromal cells. elongated nuclear morphology of these cells suggested that they were fibroblasts, although detailed analysis for their identification was not performed.
Expression of IGF-I mRNA in non-malignant breast tissue
dilatation of the terminal ductule (blunt duct adenosis). Figure 5 is an example of malignant tissue with stromal cells around both malignant cells and benign ductal components expressing IGF-II mRNA. Thus, we found IGF-II mRNA expression around stromal cells adjacent to both normal and malignant epithelial cells. Furthermore, we have seen one case of ductal carcinoma
As shown in Figure 3, IGF-I mRNA hybridization signal was localized over stromal cells (probably fibroblasts) of normal lobule (Figs. 3a and 3b) and fibroblasts in tubular adenomas and fibroadenomas (Fig. 3c). These benign diseases are thought to represent the non-malignant expansion of normal lobutar structures.
Expression of IGF-H mRNA in human breast cancer We have also preformed a limited number of IGF-II in-situ hybridizations. As shown in Figure 4, IGF-II mRNA is expressed in normal stromal cells in this near normal lobule with cystic
Figure 5. Expressionof IGF-IImRNA in infiltratingductal carcinoma (bright field photomicrograph, 100X). Hybridization signals are localized over stromal cells around normal (N) as well as malignantepithelial cells (T).
36
S Paik
Figure 6. Expression of IGF-II mRNA in this case of intraductal carcinoma, solid type, is localized over
the epithelial component (darkfield photomicrograph). Fibrotic stromal cells around the duct failed to hybridize with IGF-II probe. in-situ in which malignant epithelial cells themselves seem to express IGF-II mRNA (Figure 6).
Discussion IGF-I is regarded as the adult IGF species while IGF-II is preferentially expressed by immature organogenic cells during development [4]. In this study, IGF-II mRNA was localized in the stromal cells around malignant neoplastic cells while IGFI mRNA was seen exclusively associated with normal or benign breast epithelial cells. Such data suggests that the epithelial cells may modulate stromal cell expression of the IGFs, as represented schematically in Figure 7. It is possible that malignant epithelial cells produce factors which provide selective advantage for IGF-II producing fibroblasts. However, whether loss of expression of stromal IGF-I mRNA around malignant epithelial cells is due to inhibition of transcription or to the clonal selection of a specific population of stromal cells needs to be
tested. Cullen et al. have demonstrated that fibrobtasts obtained from malignant vs. benign breast biopsy specimens differed in their IGF expression [3]. Since differences in IGF expression appear to remain even after the fibroblast has been separated from its associated epithelial cell, clonal modulation seems to be a more reasonable interpretation of the data. What is the effect of IGF class switching on malignant epithelial cells? Expression of IGF-II rather than IGF-I should not offer any growth advantage to the tumor cells with type I IGF receptor, since IGF-I and IGF-II are approximately equipotent [5]. Effects other than mitogenesis, such as motility, may be mediated through an IGF-II/type II IGF receptor interaction [6], although the relevance of these findings to breast cancer has not been demonstrated. More likely, the stromal cells associated with malignant epithelial cells have a more fetal phenotype and IGF-II expression is merely a marker for this phenotype. These "fetal" stromal cells may offer growth advantages that are not related to IGF
IGF mRNA expression in breast tissue
37
IGF-II producer IGF-II producer
IGF-I p r o d u c e ~
Clonal modulatio of fibroblasts by cancer cells
Normal
Cancer
Figure 7. Schematic summary of in-situ hybridization data. Both IGF-I and IGF-II mRNA are expressed in normal lobular stromal cells. In infiltrating ductal carcinoma, stromal cells around malignant epithelial cells express only IGF-II mRNA.
expression. Thus, in my view, IGF class switching is simply a marker (or end result) of epithelial-stromal celt interaction in malignancy. Modulation of stromal gene expression by cancer cells is not new. In the literature, several molecules have been described whose expression is induced by the stromal cells around breast cancer cells. These include tenascin/hexabrachion/cytotactin [7] and gelsolin [8]. Expression of tenascin has been shown to be induced by TGF[3 secreted by breast cancer cells in vitro. These data suggest that there is a dynamic interaction between epithelial and stromat components in malignant processes which may be amenable to therapeutic intervention.
Acknowledgements The author would like to dedicate this article to the late Dr. William McGuire for his past encouragement. Earlier assistance from Drs. Enzo Bard (Department of Pathology, State University of New York Health Science Center at Brooklyn),
Neal Rosen (Memorial Sloan-Kettering Hospital, New York, NY), and Douglas Yee (University of Texas Health Science Center, San Antonio, TX) is greatly appreciated.
References 1. Yee D, Paik S, Lebovic GS, Marcus RR, Favoni RF, Cullen KJ, Lippman ME, Rosen N: Analysis of insulin-like growth factor I gene expression in malignancy: evidence for a paracrine role in human breast cancer. Mol Endocrinol 3:509-517, 1989 2. Hogan B, Costantini F, Lacy E: Manipulating the Mouse Embryo - - A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, 1986, p228 3. Cullen KJ, Smith HS, Hill S, Rosen N, Lippman ME: Growth factor mRNA expression by human breast fibroblasts from benign and malignant lesions. Cancer Res 51:4978-4985, 1991 4. Paik S, Rosen N, Perdue JF, You JM, Lippman ME, Yee D: Expression of IGF-II mRNA in fetal kidney and Wilms' tumor; an in-situ hybridization study. Lab Invest 61:522-526, I989 5. Yee D, Cullen KJ, Paik S, Perdue J, ttampton B, Schwartz A, Lippman ME, Rosen N: Insulin-like growth factor II mRNA expression in human breast
38
S Paik
cancer. Cancer Res 48:6691-6696, 1988 6. Minniti CP, Kohn EC, Grubb JH, Sly WS, Oh Y, Muller HL, Rosenfetd RG, Helman LJ: The insulin-like growth factor II (IGF-II)/mannose-6-phosphate receptor mediates IGF-II-induced motility in human rhabdomyosarcoma cells. J Biol Chem 267:9000-9004, 1992 7. Natali PG, Nicotra MR, Bigotti A, Botti C, Castellani
P, Risso AM, Zardi L: Comparative analysis of the expression of the extracellular matrix protein tenascin in normal human fetal, adult, and tumor tissues. Int J Cancer 47:811-816, 1991 Chaponnier C, Gabbiani G: Gelsolin modulation in epithelial and stromal cells of mammary carcinoma. Am J Pathol 134:597-603, 1989
Expression of IGF-I and IGF-II mRNA in breast tissue Soonmyoung Paik Department of Pathology, Lombardi Cancer Center, Georgetown University, Washington DC, USA
Key words: breast tumor cells, epithelial cells, IGF-I, IGF-II, in-situ hybridization, insulin-like growth factors, messenger RNA, stromal cells
Summary Expression of IGF-I and IGF-II was studied in human breast cancer tissues by in situ hybridization. IGF-I mRNA was detected only in stromal cells adjacent to normal breast epithelial cells. Stromal cells associated with the tumor cells did not contain IGF-I, nor did malignant or benign breast epithelial ceils. In contrast, IGF-II mRNA was found in both the malignant epithelial cells and their adjacent stromal cells. These data imply that stromal cells associated with breast epithelium may switch expression from IGF-I to IGF-II during breast cancer evolution. This appearance of IGF-II expression may identify cancerassociated stromal cells that have a fetal phenotype.
Introduction Although earlier in-vitro studies suggested that IGF-I might be an estrogen inducible autocrine growth factor for breast cancer cells, measurement of IGF-I mRNA levels in homogenized breast tissue specimens using RNase protection assay showed an unexpected result - - there was decreased expression of IGF-I mRNA in malignant tissues when compared to adjacent normal tissues [1]. This data, combined with the failure to demonstrate IGF-I expression in breast cancer cell lines, suggested that IGF-I was unlikely to be produced by malignant epithelial cells. Several alternative hypotheses were developed to explain the detection of IGF-I in breast cancer tissues: 1) IGF-I mRNA is expressed by normal roam-
mary epithelial cells but not by malignant cells; or 2) stromal cells in the normal lobule produce IGF-I while stromal cells associated with the cancer do not. In order to distinguish between these two possibilities, we have used in-situ hybridization to study the expression of IGF-I mRNA in paraffin-embedded sections.
Methods Probes
The IGF-I cDNA was kindly provided by Ken Gabby (Baylor College of Medicine, Houston, TX). The 540bp PstI-BamHI fragment of IGF-IA cDNA was subcloned into pGem4 (Promega) after
Addressfor offprints: SoonmyoungPaik, M.D., Room S-165-A,LombardiCancerCenter, 3800 ReservoirRd NW, Washington DC 20007, USA
32
S Paik
removal of 21bp of the 5' poly(G) tail. IGF-II cDNA was a kind gift from Graeme Bell (Howard Hughes Institute, Chicago, IL). The 854bp PstIPstI fragment from the coding region was cloned into pGem4. Labeled antisense RNA was transcribed using T7 polymerase according to instructions of the manufacturer. Labeled sense RNA was transcribed using SP6 polymerase. 35SUTP (NEN, Wilmington, DE) was triple concentrated by lyophilization before synthesis of the probe.
Hybridization protocol Previously described methods [2] were used except for modifications of the pretreatment step for paraffin-embedded sections. All pre-ribonuclease (RNase) washing steps were carried out using autoclaved polypropylene staining dishes (Miles Lab). Sections were deparaffinized and rehydrated through alcohol series. Basic proteins were removed by incubating 20 minutes in 0.2M HCI. After washing in distilled water for 5 minutes, sections were denatured by incubating 30 minutes at 70°C in 2xSSC. Sections were then digested with 0.5 /ag/ml proteinase K in 20mM Tris-HC1/2mM CaC12 for 3 minutes at room temperature. Sections were then incubated 30 minutes in 0.1M Tris-0.1M Glycine, washed twice for 3 minutes each in Dutbecco's PBS, and acetylated for 10 minutes in fresh acetic anhydride diluted 1:400 in 0.1M triethanolamine (pH 8.0). Sections were then washed in Dulbecco's PBS for 5 minutes and dehydrated through 30%, 60%, 80%, 95%, and 100% ethanol for 2 minutes each and air dried for 1 hour. Hybridization mix contained 50% formamide, 10% dextran sulfate, 2xSSC, 2mg/ml bovine serum albumin, lmg/ml salmon sperm DNA, lmg/ml yeast tRNA, 50raM DTT, and 5X10 7 cpm/ml of probe. Probe mix was prepared and kept in a 50°C water bath. Sections were covered with 100 to 200 pl of hybridization mix by careful spreading with a pipette tip with-
out touching the section. No coverslip was used to prevent tissue detachment. Hybridization time was 4 hours. Sections were transferred to 50% formamide/2xSSC at 52°C to wash off hybridization mix, washed in 50% formamide/2xSSC with 10raM DTT at 52°C for 5 minutes with continuous vertical washing, then washed for 20 minutes in a new batch of same solution, washed 4 times 1 minute in 2xSSC at room temperature, incubated in 20~g/ml RNase A prepared in 2xSSC for 30 minutes at 37°C, and rinsed twice 1 minute each in 2xSSC. Final wash was done in 50% formamide/2xSSC without DTT at 60°C for 30 minutes with continuous agitation. After washing 1 minute in 2xSSC, sections were dehydrated through 70%, 80%, and 95% ethanol for 1 minute each and air dried for 1 hour. Sections were then exposed to Hyperfilm-beta max (Amersham) for 3 days to check the completeness of washing, and if necessary rehydrated and washed again in 0. l xSSC for 30 minutes with continuous agitation and dehydrated. Sections were then exposed to NTB-2 emulsion for 3 weeks, developed for 2.5 minutes in Kodak developer, fixed for 3 minutes in Kodak fixer, washed for 5 minutes in distilled water, and air dried. Sections were then stained with hematoxylin and eosin, dehydrated, coverslipped, and examined under a microscope. After examination and photography, cover slips were removed and the sections were re-exposed for 3 weeks to Hyperfilm-beta max. Emulsion exposure and film exposure were then compared to rule out possible artifacts in emulsion exposure.
Optimization of" the assay conditions The assay was optimized using cell lines known to express only one of the IGF species: cells expressing only IGF-I mRNA (CHP-100, American Type Culture Collection, Rockville, MD), and cells which express only IGF-II mRNA (HepG2, ATCC). Time course of the hybridization reaction showed maximum signal to noise ratio
IGF mRNA expression in breast tissue
33
hybridization. Note strong black signal over CHP-100 cells and no signal over HepG2 cells. A case of tubular adenoma with known m R N A expression by RNase protection assay was fixed overnight in 10% formalin and embedded in paraffin. This section was used as a positive control for subsequent assays.
Figure 1. Example of film exposure for IGF-I in-situ hybridization. IGF-I expressing CHP100 cells show strong hybridization signal whereas IGF-II expressing control cells (HepG2) show no signal. after 4 hours at 52°C. Background signal increased exponentially after 4 hours hybridization without increase in specific hybridization signal. Washing conditions were optimized by exposing the sections to the beta-sensitive film (Hyperfilm beta max) before emulsion exposure. Figure 1 is an example of film exposure of positive and negative control cell lines for IGF-I in-situ
Results Expression of IGF-I mRNA in human breast cancer None of the sections from 10 breast cancer cases (without normal lobules in the section) showed significant IGF-I expression. However, in two cases in which a normal lobular component was associated with the breast cancer, expression of IGF-I mRNA was detected limited to stromal cells in the non-malignant area (Figure 2). The
Figure 2. Expression of IGF-I mRNA in infiltrating ductal carcinoma of the breast (100X). Note hybridization signal over stromal cells in non-malignant areas (N) whereas there are no signals over either malignant epithelial cells or stromal cells around them (T).
34
S Paik
Figure 3. A) Expression of IGF-I mRNA in normal lobule (100X). Hybridization signals are confined over a subset of stromal cells (S) in the Iobule. B) Positive cells are more abundant in edematous area (E), C) Expression of IGF-I mRNA in fibroadenoma (100X). All hybridization signals are confined to stromal component (S) and not epithelial cells (E).
IGF mRNA expression in breast tissue
35
Figure 4. Expressionof IGF-II mRNA in non-malignantbreast tissue, in this case with blunt duct adenosis (dark field photomicrograph, 100X magnification). Hybridizationsignals are over stromal cells. elongated nuclear morphology of these cells suggested that they were fibroblasts, although detailed analysis for their identification was not performed.
Expression of IGF-I mRNA in non-malignant breast tissue
dilatation of the terminal ductule (blunt duct adenosis). Figure 5 is an example of malignant tissue with stromal cells around both malignant cells and benign ductal components expressing IGF-II mRNA. Thus, we found IGF-II mRNA expression around stromal cells adjacent to both normal and malignant epithelial cells. Furthermore, we have seen one case of ductal carcinoma
As shown in Figure 3, IGF-I mRNA hybridization signal was localized over stromal cells (probably fibroblasts) of normal lobule (Figs. 3a and 3b) and fibroblasts in tubular adenomas and fibroadenomas (Fig. 3c). These benign diseases are thought to represent the non-malignant expansion of normal lobutar structures.
Expression of IGF-H mRNA in human breast cancer We have also preformed a limited number of IGF-II in-situ hybridizations. As shown in Figure 4, IGF-II mRNA is expressed in normal stromal cells in this near normal lobule with cystic
Figure 5. Expressionof IGF-IImRNA in infiltratingductal carcinoma (bright field photomicrograph, 100X). Hybridization signals are localized over stromal cells around normal (N) as well as malignantepithelial cells (T).
36
S Paik
Figure 6. Expression of IGF-II mRNA in this case of intraductal carcinoma, solid type, is localized over
the epithelial component (darkfield photomicrograph). Fibrotic stromal cells around the duct failed to hybridize with IGF-II probe. in-situ in which malignant epithelial cells themselves seem to express IGF-II mRNA (Figure 6).
Discussion IGF-I is regarded as the adult IGF species while IGF-II is preferentially expressed by immature organogenic cells during development [4]. In this study, IGF-II mRNA was localized in the stromal cells around malignant neoplastic cells while IGFI mRNA was seen exclusively associated with normal or benign breast epithelial cells. Such data suggests that the epithelial cells may modulate stromal cell expression of the IGFs, as represented schematically in Figure 7. It is possible that malignant epithelial cells produce factors which provide selective advantage for IGF-II producing fibroblasts. However, whether loss of expression of stromal IGF-I mRNA around malignant epithelial cells is due to inhibition of transcription or to the clonal selection of a specific population of stromal cells needs to be
tested. Cullen et al. have demonstrated that fibrobtasts obtained from malignant vs. benign breast biopsy specimens differed in their IGF expression [3]. Since differences in IGF expression appear to remain even after the fibroblast has been separated from its associated epithelial cell, clonal modulation seems to be a more reasonable interpretation of the data. What is the effect of IGF class switching on malignant epithelial cells? Expression of IGF-II rather than IGF-I should not offer any growth advantage to the tumor cells with type I IGF receptor, since IGF-I and IGF-II are approximately equipotent [5]. Effects other than mitogenesis, such as motility, may be mediated through an IGF-II/type II IGF receptor interaction [6], although the relevance of these findings to breast cancer has not been demonstrated. More likely, the stromal cells associated with malignant epithelial cells have a more fetal phenotype and IGF-II expression is merely a marker for this phenotype. These "fetal" stromal cells may offer growth advantages that are not related to IGF
IGF mRNA expression in breast tissue
37
IGF-II producer IGF-II producer
IGF-I p r o d u c e ~
Clonal modulatio of fibroblasts by cancer cells
Normal
Cancer
Figure 7. Schematic summary of in-situ hybridization data. Both IGF-I and IGF-II mRNA are expressed in normal lobular stromal cells. In infiltrating ductal carcinoma, stromal cells around malignant epithelial cells express only IGF-II mRNA.
expression. Thus, in my view, IGF class switching is simply a marker (or end result) of epithelial-stromal celt interaction in malignancy. Modulation of stromal gene expression by cancer cells is not new. In the literature, several molecules have been described whose expression is induced by the stromal cells around breast cancer cells. These include tenascin/hexabrachion/cytotactin [7] and gelsolin [8]. Expression of tenascin has been shown to be induced by TGF[3 secreted by breast cancer cells in vitro. These data suggest that there is a dynamic interaction between epithelial and stromat components in malignant processes which may be amenable to therapeutic intervention.
Acknowledgements The author would like to dedicate this article to the late Dr. William McGuire for his past encouragement. Earlier assistance from Drs. Enzo Bard (Department of Pathology, State University of New York Health Science Center at Brooklyn),
Neal Rosen (Memorial Sloan-Kettering Hospital, New York, NY), and Douglas Yee (University of Texas Health Science Center, San Antonio, TX) is greatly appreciated.
References 1. Yee D, Paik S, Lebovic GS, Marcus RR, Favoni RF, Cullen KJ, Lippman ME, Rosen N: Analysis of insulin-like growth factor I gene expression in malignancy: evidence for a paracrine role in human breast cancer. Mol Endocrinol 3:509-517, 1989 2. Hogan B, Costantini F, Lacy E: Manipulating the Mouse Embryo - - A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, 1986, p228 3. Cullen KJ, Smith HS, Hill S, Rosen N, Lippman ME: Growth factor mRNA expression by human breast fibroblasts from benign and malignant lesions. Cancer Res 51:4978-4985, 1991 4. Paik S, Rosen N, Perdue JF, You JM, Lippman ME, Yee D: Expression of IGF-II mRNA in fetal kidney and Wilms' tumor; an in-situ hybridization study. Lab Invest 61:522-526, I989 5. Yee D, Cullen KJ, Paik S, Perdue J, ttampton B, Schwartz A, Lippman ME, Rosen N: Insulin-like growth factor II mRNA expression in human breast
38
S Paik
cancer. Cancer Res 48:6691-6696, 1988 6. Minniti CP, Kohn EC, Grubb JH, Sly WS, Oh Y, Muller HL, Rosenfetd RG, Helman LJ: The insulin-like growth factor II (IGF-II)/mannose-6-phosphate receptor mediates IGF-II-induced motility in human rhabdomyosarcoma cells. J Biol Chem 267:9000-9004, 1992 7. Natali PG, Nicotra MR, Bigotti A, Botti C, Castellani
P, Risso AM, Zardi L: Comparative analysis of the expression of the extracellular matrix protein tenascin in normal human fetal, adult, and tumor tissues. Int J Cancer 47:811-816, 1991 Chaponnier C, Gabbiani G: Gelsolin modulation in epithelial and stromal cells of mammary carcinoma. Am J Pathol 134:597-603, 1989
