| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
Center for Cancer Biology and Nutrition, Institute of Biosciences and Technology, Texas A&M University System Health Science Center [X. W., C. J., F. W., C. Y., W. L. M.], and Graduate School of Biomedical Sciences, The University of Texas Health Science Center at Houston [X. W., C. Y.], Houston, Texas 77030
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-actin. The other (DTS2) exhibited SMC -actin and an SMC-like morphology in vitro. Both subtypes expressed FGF7 and equally low levels of FGFR2IIIc mRNA, whereas fibroblast growth factor receptor (FGFR) 1 predominated in DTS1 cells. DTS1 cells also expressed FGF10 and no detectable FGFR3, whereas the absence of FGF10 and presence of FGFR3 distinguished DTS2 cells. Epithelial cell-derived FGF9 bound to FGFR and stimulated growth of specifically FGFR3-positive DTS2 cells, not the FGFR3-negative DTS1 cells. These results demonstrate stromal cell heterogeneity in signal reception of FGF from epithelium. This correlated with potential heterogeneity in the response back to epithelial cells. Epithelium-dependent control of a stromal cell phenotype within a tumor may be a determinant of whether tumors remain in nonmalignant homeostasis or progress to malignancy. |
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
To study changes in mutual communication mediated by the FGF signaling system between stromal and epithelial cells during malignant progression, we have used a rare rat prostate adenocarcinosarcoma (Dunning R3327PAP) in which stromal and epithelial compartments have evolved as a single mutually interdependent nonmalignant transplantable unit (8 , 11, 12, 13, 14, 15) . Tumors implanted under the skin grow slowly without severe consequence to the host to near the weight of the male host, while maintaining the same ratio of epithelium to stroma. On castration of hosts or reimplantation of isolated epithelial cells in absence of stromal cells, epithelial cells progress to malignancy that can metastasize and kill the host. However, if isolated stromal cells are implanted together with epithelial cells at the level present in the parent nonmalignant tumor, homeostatic balance between growth and differentiation between the two compartments is restored with no progression of epithelial cells to malignancy (8 , 11) . In contrast to epithelial cells, stromal cells did not survive when implanted alone, but required the presence of the epithelial cells. This confirmed that, in addition to directional instructive signals from stroma to epithelium, signals from epithelium to stroma complete an overall homeostasis-promoting dialogue between compartments. FGF7 and FGF10, and the specific splice variant FGFR2IIIb complexed with cell type and receptor isotype-specific heparan sulfate have been well-characterized as a paired signal and receptor that mediates directionally specific communication from stromal to epithelial cells (8 , 11 , 16) . FGF7 and FGF10, expressed only in stromal cells, are absent in epithelial cells, and specifically bind and activate a complex of heparan sulfate and FGFR2IIIb that is expressed only in epithelial cells (8 , 11 , 16, 17, 18) . The loss of FGFR2IIIb is a hallmark of malignant stroma-independent epithelial cells from a variety of parenchymal tissues including prostate (8 , 11 , 19, 20, 21) . Restoration of the FGFR2 kinase to malignant epithelial cells devoid of it restores elements of homeostasis associated with nonmalignant cells such as limited cell proliferation, apoptosis, differentiation, and response to stroma (11 , 20, 21, 22) . Despite the critical impact of stroma on the character of epithelium, little is known about the specific character of the stromal compartment and signals that flow from epithelium in premalignant tumors that are in relatively benign homeostasis. In this report, we performed a clonal analysis of stromal cells derived from the Dunning R3327PAP adenocarcinosarcoma, and characterized them according to cytoskeletal markers and expression of signaling polypeptides and receptor isotypes within the FGF family. The results revealed two distinct subtypes, one had an undifferentiated fibroblast-like character, whereas the other exhibited SMC-like properties. The two morphological subtypes exhibited distinct expression patterns with respect to FGF7, FGF10, FGFR1, FGFR2IIIc, and FGFR3. The results suggest that stromal cell subtypes respond to and communicate back to epithelial cells differently. The results suggest a specific epithelial to stromal cell part of a two-way stromal-epithelial cell dialogue in premalignant slow-growing differentiated tumors of which the disruption may contribute to progression of epithelial cells to malignancy.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
5 x 106 cells harvested from monolayers of established stock cultures were serially diluted to 10 cells/ml in medium containing 10% FBS, and 0.10 ml/well was distributed among wells of 98-well plates. To ensure clonal cultures were derived from a single cell, wells were observed continuously after inoculation. About 50% of wells gave rise to colonies, and cultures were established subsequently from wells containing single colonies in wells containing only single cells at inoculation. Sequential transfer of clonal cultures to 24-well, 12-well, and 6-well plates, and then 25-cm2 flasks established stock cultures for experiments.
Immunocytochemical Analysis of Cultured Cells.
Cells were seeded into vitrogen-coated polystyrene chambers with surface area 2 cm2 (Tissue Culture Treated Glass Slide, Franklin Lakes, NJ) at a density of 5 x 104 to 5 x 105 cells/cm2 and fixed with glutaraldehyde-acetone (Sigma, St. Louis, MO) at -20°C for 15 min after 24-h incubation. A minor modification of manufacturer’s instructions from -SMOOTH MUSCLE ACTIN kit (IMMH2; Sigma Chemical Co.) was used as follows using the amount of primary affinity-purified monoclonal antibody in the text. Endogenous peroxidase was quenched by incubation with 3% hydrogen peroxide for 10 min before introduction of the primary antibodies. Slides were incubated with biotinylated goat antimouse immunoglobulin at a 1:5 dilution followed by treatment with the extra-avidin peroxidase reagent at a 1:20 dilution from Sigma kit IMMH2. Antigen was visualized by addition of freshly prepared hydrogen peroxide and the chromogen 3-amino-9-ethyl-carbazole (Sigma kit IMMH2). Negative controls were established by omission of the primary antibody. Primary antibodies used reacted against human, mouse, rat, and chicken antigen. Anti-SMC -actin (Sigma) was prepared against human -actin, and reacted with normal and neoplastic smooth muscle, and myoepithelial cells. Antikeratin (pancytokeratin; Sigma) was a mixture of mouse monoclonal antibodies A53-B/A2 (MCF7 cells), CY-90 (A431 and MCF7 cells), C-11 and KS-1A3 (A431 cells), M20 (MCF7 cells), and PCK-26 (human epidermis) prepared against keratin from the human sources in parentheses. The pancytokeratin reacted with keratin subtypes 1, 4–6, 8, 10, 13, 18, and 19. Antivimentin (Sigma) was mouse hybridoma clone V9 against pig lens vimentin.
Immunohistochemical Analysis of Tissues.
Fresh tissues of NPS or tumors were fixed with HISTOCHOICE tissue fixative MB (AMRESCO, Solon, OH), dehydrated by sequential ethanol gradient from 70 to 100% and paraffin-embedded after coating with MEGA-CASSETTE from Sakura Fineteck USA, Inc. (Torrance, CA). Tissue sections of 5 µm were prepared, deparaffinated in xylene, and rehydrated stepwise from ethanol to PBS, and then treated with primary and secondary antibodies as described above for analysis of cultured cells.
Immunoblot Analysis of Cell and Tissue Extracts.
Cultured cell monolayers were lysed with RIPA buffer (1x PBS, 1% NP40, 0.5% sodium deoxycholate, and 0.1% SDS) containing protease inhibitors. The crude lysate was clarified by centrifugation (14,000 rpm) for 30 min at 4°C. Protein was determined with the bicinchoninic acid Protein Assay (Pierce, Rockford, IL). Supernatants in the amounts indicated in the text were subjected to analysis by 10% SDS-PAGE, electroblotted to nitrocellulose membranes, and then analyzed with the primary antibodies described for histochemical analysis at a 1:700 dilution. Monoclonal anti-ß-actin (clone AC-15; Sigma) was used for immunoblot at a dilution of 1:2000 as an extra loading control in some cases. The immunoblots were detected with alkaline phosphatase-conjugated secondary antibody (Sigma) and visualized using 5-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium (Sigma).
Analysis of mRNA Expression.
For RT-PCR analyses, total RNA was extracted from monolayer cell cultures with the RNeasy Mini kit (Qiagen, Valencia, CA) and reverse transcription was carried out with SuperScript II (Life Technologies, Inc., Grand Island, NY) and random primers according to protocols provided by the manufacturer. The PCR was carried out for 40 cycles at 94°C for 1 min, 55 or 60°C for 1 min, and 72°C for 1 min using TaqDNA Polymerase (Promega, Madison, WI) and the specific primers listed in Table 1
. RT-PCR products were analyzed as described in the text on 1–1.5% agarose gels.
|
-[32P]UTP during transcription by either T3 or T7 RNA polymerase, and 25 µg of total RNA was used in each reaction. Protection probes for FGFR1, 2, and 3 were against the IIIc isoforms, and spanned splice junctions between immunoglobulin modules II and III that enabled discrimination between IIIb or IIIc isoforms. The FGFR1 cDNA template was prepared from RNA from DTS cells prepared by RT-PCR using primers (sense P5R1-681 5'-CGGGATCCTACACCTGCATCGTGGAGAACGA-3' and antisense P3R1-987 5'-GCGAATTCCGTAGAAGCACCTCCATTTC-3') to yield a probe of 305 nt. The fragment was cloned into pBluescript SK vector between BamHI and EcoRI sites. A 214-nt FGFR2 probe was prepared from rat liver cDNA template (24)
. The 502-nt FGFR3 probe was prepared from rat brain cDNA template as described.7
Probes for rat FGF7 and FGF10 were prepared from PCR-amplified cDNA fragments from DTS cells using the primers for RT-PCR analysis (Table 1)
. PCR products were cloned into pBluescript SK vector at SmaI site after treatment with Klenow fragment (Life Technologies, Inc). The rat ß-actin probe was transcribed by T7 RNA polymerase using the pTRI-ß-actin-125-rat antisense control template from Ambion that contained a 139-bp cDNA fragment of rat ß-actin. Each labeled antisense RNA probe was hybridized with the samples indicated in the text at 68°C for 11 min and incubated in the presence of Ambion RNase A1/T1 at 37°C for 30 min. Protected fragments were separated on 5% polyacrylamide sequencing gels and subjected to autoradiography.
Radiolabeled Receptor Binding Assays.
FGF1 was purified from bovine brain and FGF2 from human recombinant from Upstate Biotechnology (Lake Placid, NY; Ref. 22
). Rat recombinant FGF7 was prepared as described (25)
. Rat recombinant FGF9 was prepared as follows. An FGF9 cDNA coding for 165 residues of FGF9 beginning with serine 34 fused with glutathione S-transferase with a thrombin cut site at the NH2 terminus was expressed in bacteria as a Mr 47,000 product according to the generic procedure described in our laboratory (25)
. The fusion product was purified by glutathione-Sepharose 4B affinity chromatography followed by removal of the fusion partner upstream of a glycine one residue NH2-terminal to FGF9 serine 34 with thrombin treatment. The recombinant GlySer (34)
-FGF9 was recovered and purified by heparin affinity column using fast protein liquid chromatography. The preparation yielded a single band at Mr 21,000 on SDS-PAGE that was verified by NH2-terminal sequencing. Details of the preparative procedure, and special problems associated with activity and quality control of FGF9 will be reported separately.7
General problems in preparation, iodination, and quality control of radiolabeled FGFs have been described previously (25, 26, 27)
.
For binding assays, cells were seeded at 0.5 x 105 cells/cm2 into six-well plates containing 3 ml of RD medium supplemented with 2% FBS and incubated overnight. The medium was then removed, and 1 ml of serum-free medium containing 2 ng of iodinated FGF was added to each well. After 1-h incubation at 37°C, the medium was removed; the cells were rinsed with PBS, then 250 µg/ml heparin in PBS, and incubated with 1 ml of PBS containing 1 mM disuccinimidyl suberate for 15 min under gentle shaking and lysed in Triton X-100. The lysate was clarified for 30 min by centrifugation (14, 000 rpm) at 4°C and mixed with sample buffer for SDS-PAGE followed by autoradiography (9)
.
Measurement of DNA Synthesis by [3H]Thymidine Incorporation.
Optimum culture densities and conditions were established for each cell type that yielded statistically significant levels of thymidine incorporation per cell in response to the indicated external FGF over background incorporation. Cells were seeded into 24-well plates containing 1 ml of RD medium and 5% FBS for 18 h. DTS1 cells were seeded at a density of 2 x 104 and 6.5 x 104 cells/cm2 for assay of effect of FGF1 and FGF9, respectively. DTS2 cells were introduced at 6.5 x 103 cells/cm2 for assay of both factors. Medium in DTS2 cultures was replaced with serum-free medium and that for DTS1 cultures with medium containing 0.5% FBS for 48 h before assay. FGF was added as indicated in the text for 18–20 h, and then [3H]thymidine (DuPont NEN Life Science Products, Boston, MA) was added to a final concentration of 0.31 pM (specific activity 6700 Ci/mol). After 4 h, the cells were rinsed sequentially with PBS, cold 10% trichloroacetic acid, and PBS before solubilization in 200 µl of 0.50 M NaOH. The [3H]thymidine incorporated into trichloroacetic acid-insoluble material was measured by liquid scintillation. Cell number at the time of harvest was determined by direct Coulter count in replicate cultures. Incorporation in dpm was normalized to the number of each cell type.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The remaining 4 (5%) clonal lines (DTS2) exhibited an SMC-like character based on display of both vimentin and SMC actin. Immunoblot analysis of cell extracts confirmed expression of the indicated cytoskeletal antigens revealed by immunohistochemistry of clonal cultures representative of the two subtypes (Fig. 2)
.
|
|
-actin-staining cells that tightly abut on and surround the epithelial cells that surround the lumen (Ref. 28
; Fig. 3A
). Less intimately associated with the epithelial cells is the diverse minority of nonparenchymal cells in the tissue matrix. Similar to normal prostate, morphologically differentiated two-compartment DT tumors exhibit lumen-like structures formed by epithelial cells that are smaller, more abundant, and much less organized and heterogeneous relative to normal prostate (15)
. Nevertheless, the DT tumors exhibit SMC actin-staining cells lining the lumen on the matrix or basal side of the lumenal epithelial cells (Fig. 3A)
. The SMC actin-staining cells were notably less organized, sometimes overlapping or multilayered, and exhibited numerous gaps in the continuous ring surrounding the epithelial cells that was evident in normal prostate. Characteristic of the DT tumors is a much more extensive and heterogeneous stromal matrix than normal tissue that can comprise up to 40% of the tumor volume (8
, 11
, 15)
. Similar to normal tissue, cells that lie in the more extensive stromal matrix appear to be mostly fibroblast-like cells with no apparent SMC actin expression and are dispersed in the matrix relative to lumenal-associated SMC actin-staining cells. Immunoblot analysis confirmed that the total expression of -actin throughout normal and DT tumor tissue was equal (Fig. 3B)
in contrast with DT tumor-derived malignant tumors (AT) that exhibited no distinct stromal compartment (Ref. 8
; Fig. 3A
).
|
-actin-positive cells in vitro from the onset of culture or a gradual loss of the SMC-like property with subsequent passage in culture. Similar to stromal cultures from normal prostate (Fig. 4
; Fig. 5A
), cells with both the morphology (Fig. 4)
and the SMC -actin expression (Fig. 5B)
of the clonal line of DTS2 cells predominated in early passage cultures from DT tumors. This was apparent as soon as selective culture conditions yielded stromal cultures free of epithelial cells in sufficient amounts for analysis. However, by passage 70, cells expressing SMC -actin were sparse, and the cultures were dominated by the DTS1 phenotype. The reciprocal relationship between ratio of the two subtypes of stromal cells according to morphology (Fig. 4)
and passage in vitro was quantitated and summarized in Table 2
. The results show that, similar to normal prostate, SMC -actin-expressing DTS2 cells represent the predominant stromal cell type from DT tumors, but the property is lost with prolonged passage in vitro (Fig. 5, B and C)
. Despite years of s.c. passage in syngenic hosts in vivo, the two-compartment DT tumors gave rise predominately to the DTS2 cell type in culture that appears similar to the SMC-like cells from NPS.
|
|
|
. Analysis of expression levels by RNase protection (RPA) in two different clonal cultures of each cell type confirmed that FGF7 and FGF10 mRNA were expressed significantly above trace levels detected by RT-PCR (Fig. 6B)
. Both DTS1 and DTS2 expressed FGF7 at near equal levels, whereas expression of FGF7 in DTS2 cells was slightly higher. Levels of FGF7 and FGF10 mRNAs in DTS1 cells were similar. These results show that although FGF7 and FGF10 both interact in vitro with epithelial cell FGFR2IIIb in complex with epithelial cell heparan sulfate (16
, 23
, 27)
, the expression of FGF7 and FGF10 is partitioned between different stromal cell types.
|
. Expression of FGFR3 was observed in the DTS2 cultures, as well as in the parent HDTS cultures and lower-passage stromal cultures (LDTS). The FGFR3 signal was barely detectable in most and undetectable in some clonal DTS1 lines. Analysis by RPA that requires higher levels of mRNA than the RT-PCR analysis using a probe comprised of common and IIIc isoform-specific sequence revealed that only the IIIc isoforms of FGFR1 and FGFR2 were expressed in both DTS1 and DTS2 clonal lines (Fig. 7B)
. Among the clonal lines examined, FGFR1IIIc was expressed in DTS1 cells at least 20 times higher than that in DTS2 cells, and in some DTS2 clonal lines, it was barely detectable. Expression of FGFR2IIIc that was apparently equal in both clonal lines was markedly lower than FGFR1 in the DTS1 lines and FGFR3 in the DTS2 lines (Fig. 7B)
. Both the IIIc and IIIb isoforms of FGFR3 at a ratio of 5:1, respectively, were specifically expressed in the SMC -actin-staining DTS2 cells (Fig. 7B)
. Although trace amounts were revealed by RT-PCR, the RPA suggested that FGFR3 mRNA expression was insignificant in the DTS1 lines. It was noteworthy that clonal variation of the intensity of the total FGFR3 signal among DTS2 lines was inversely proportional to the expression of FGFR1 mRNA.
|
-Actin-positive Stromal Cells Expressing FGFR3.
. FGF9 failed to bind to DTE cells that express only FGFR2IIIb (8
, 14 , 17)
, whereas FGFR2IIIb-specific FGF7 bound only to DTE cells (Fig. 8B)
. These results show that only SMC -actin-positive, FGFR3-expressing DTS2 cells exhibit high affinity receptor sites for FGF9 and, therefore, are likely to be selectively responsive to FGF9 (Fig. 8B)
.
|
-Actin-positive DTS2 Stromal Cells.
. FGF9 specifically stimulated DNA synthesis of the FGFR3-positive DTS2 cells (Fig. 9A)
, whereas FGF1 stimulated DNA synthesis of both DTS1 and DTS2 stromal lines (Fig. 9B)
. FGF9 stimulated proliferation of the DTS2 cell population by 2-fold with no effect on DTS1 under the same conditions (Fig. 9C)
. Although the intrinsic rate of DNA synthesis and cell population growth rate of DTS1 cells were markedly lower than the DTS2 cells, the response to FGF1 (Fig. 9B)
and FBS (Fig. 9C)
indicated that the proliferative response of DTS1 cells was not generally defective relative to DTS2 cells.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-actin expression associated with the DTS2 subtype is similar to the transitional myofibroblast described by Powell et al. (29)
and Tuxhorn et al. (30)
. Whether DTS2 cells arise by dedifferentiation of mature SMCs or are a common precursor to both, are capable of differentiating into mature nonproliferating SMCs or dedifferentiate to DTS1 cells, is under investigation. The loss of cells with DTS2 properties and dominance of the undifferentiated fibroblast-like DTS1 cells is a consequence of prolonged culture in vitro. This may reflect the absence of epithelium and other host environmental factors required to support the DTS2 cell phenotype. We cannot rule out a contribution of the possible selective advantage of DTS1 cells conferred by the high levels of FGFR1 relative to DTS2 cells and a potential proliferative autocrine loop with FGF2 that is deficient in DTS2 cells. However, the reduced rate of proliferation of DTS1 cells in serum-free culture relative to DTS2 cells argues against a rate-limiting effect of this autocrine loop.
Stromal FGF7 and FGF10 is Partitioned between the Two Clonal Stromal Cell Types.
Expression of FGF7 and FGF10 is limited to the stromal compartment of numerous parenchymal tissues, including prostate (17)
. The two homologous, but genetically distinct, polypeptides have similar specificity for an artificial complex of heparin and FGFR2IIIb, and both exhibit a response to androgen under the same conditions in vitro (8
, 18
, 23)
. Therefore, both have been proposed as potentially redundant andromedins that mediate androgen-responsive directional communication from stroma to epithelium. However, FGF7 and FGF10 are dramatically different with respect to expression pattern in vivo and impact of ablation in mice (31, 32, 33)
. Whereas FGF7 is expressed diffusely throughout the mesenchyme in most organs, expression of FGF10 is generally present in mesenchymal cells associated with growing epithelia and is required for prostatic development (33
, 34)
. Relative to FGF7, FGF10 mRNA levels are low in the adult prostate (23
, 35) . Consistent with these observations, we showed here that expression of FGF10 is stromal cell type-specific and absent from the SMC-like cells that comprise the majority of normal prostate and two-compartment premalignant tumors that are in benign homeostasis. DT tumors exhibit an elevated level of FGF10 expression relative to normal prostate, which is consistent with their more extensive and heterogeneous stromal matrix (8
, 11 , 15)
, and an initially higher content of FGF10-expressing DTS1 type cells both in early and late passage cultures than normal prostate (Table 2)
. It remains unclear whether these differences impact overall stromal-epithelial cell homeostasis mediated via FGFR2IIIb, because FGF7 appears present in both stromal cell types at equal levels. However, a hallmark of the FGF family is that expression at both the mRNA and protein levels rarely is the rate-limiting factor for FGF signaling in tissues (17)
. FGF- and FGFR-specific heparan sulfate limit both access of FGF to the FGFR signaling complex and the specificity of the oligomeric FGFR signaling complex (9
, 16
, 17
, 25
, 36)
. We have shown previously that FGF7 and FGF10 differ considerably in affinity for heparin and natural matrix heparan sulfate (23
, 25) . Heparan sulfate has the potential to confer specificity on the two signals through pericellular matrix-mediated differential access to epithelial FGFR2IIIb (23)
or formation of factor-specific binary complexes of heparan sulfate and the FGFR2IIIb ectodomain (16
, 36)
.
Epithelial Cell-derived FGF9 May Support the SMC-like Stromal Cell Population via FGFR3.
In the two-compartment, nonmalignant model tumor system studied here, the expression of FGF9 and FGFR3 is partitioned between the epithelial cells (DTE) and the stromal cells (DTS), respectively.7
FGF9 mRNA cannot be detected in either the uncloned DTS or derived clonal lines DTS1 or DTS2 at trace levels using RT-PCR analysis, but is present at significant levels revealed by RPA in the DTE and malignant AT3 tumor epithelial cells that evolve from them during progression to malignancy.7
Our results show that it is specifically the SMC-like stromal cell that expresses FGFR3 and binds and responds to FGF9. Therefore, within the FGF family, FGF9 may be a signal from epithelial cells that maintains the high ratio of SMC-like cells within the stroma that is characteristic of both normal prostate and nonmalignant two-compartment prostate tumors. Also worthy of note is our finding that expression of FGFR1IIIc in the clonal stromal lines was inversely proportional to the expression of FGFR3. FGFR1 is presumed to be the major FGFR isotype for FGF2 that is expressed by both DTS1 and DTS2 cells. It has been reported that FGF2 reduced the expression of SMC -actin in myofibroblasts induced by transforming growth factor ß thereby reducing the apparent number of myofibroblasts relative to fibroblasts, independent of proliferation (37
, 38)
. A deficiency of either the epithelial FGF9 signal, or reduced expression or activity of FGFR3 compounded by cell-specific changes in FGFR1 in the stroma may tip the balance toward less-differentiated stromal cell types of which the fibroblast-like stromal DTS1 cell is a prototype.
Directional Epithelial to Stromal Communication via the FGF Family in Two-Compartment Homeostasis and Progression of Epithelial Cells to Malignancy.
Directionally specific signaling from prostate stroma to epithelium is mediated in part by stromal-specific FGF7 and FGF10 signals acting on an epithelial cell-specific complex of heparan sulfate and FGFR2IIIb kinase (8
, 11
, 16
, 18
, 23
, 27
, 36)
. Although activation of FGFR2 may support expansion of epithelial cell number under some conditions (11
, 22
, 23)
, the net effect is to promote homeostasis through growth-limiting effects such as epithelial cell apoptosis or differentiation (10
, 11
, 22)
. In premalignant two-compartment tumors, this serves to hold the epithelial compartment in check and maintains the overall tumor in a benign homeostatic state. Subversion of this homeostasis may occur by alterations in the reception end of the communication system in the epithelium or the signaling end in the stroma. Alterations in epithelial cell reception mechanisms that favor escape from stromal restraints include loss of FGFR2 and emergence of ectopic FGFR1, FGF2, and FGF5 that can act in an autocrine mode on ectopic FGFR1 in the epithelial cells (8
, 10
, 11)
. Our current results suggest changes in directional cross-talk back from epithelial to stromal cells either at the signaling end in the epithelium or the reception end in the stroma that may subvert homeostasis-promoting instructions from the stroma. Because differentiated SMCs predominate in normal prostate and nonmalignant differentiated tumors, we predict that it is most likely the SMC-like DTS2 and the FGFR3-expressing cell type supported by epithelial cell FGF9 that holds premalignant epithelial cells in homeostasis. It follows that defects in FGF9 expression or access to FGFR3-expressing stroma at the epithelial end, or defects in FGFR3 expression or activity at the stromal end may unleash premalignant epithelial cells to progress to malignancy. Conceivably, the loss of differentiated SMCs and the related DTS2 cell type either by phenotypic switching or selective pressures against them may result in a net effect or even support premalignant epithelial cell progression toward malignancy. A number of reports have shown that cultured stromal cell lines from diverse sources promote cell growth and progression toward malignancy of human prostate tumor epithelial cell lines (39
, 40)
and epithelial cells primed by genetic manipulation (41
, 42)
. This occurs without evidence of growth limitation and differentiation elicited by the homologous tumor stromal cells in the model system described here. Stromal cell origin and phenotype may underlie these differences.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Supported by Public Health Service Grants R01DK35310 from the National Institute of Diabetes and Digestive and Kidney Diseases, and R01CA59971 and U01CA84296 from the National Cancer Institute. 
2 These authors contributed equally to concept and execution of the work herein.
3 Present address: Department of Integrative Biology and Pharmacology, University of Texas Health Science Center at Houston, 6431 Fannin, Houston, TX 77030.
4 Present address: Department of Cell Biology, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030.
5 To whom requests for reprints should be addressed, at Institute of Biosciences and Technology, Texas A&M University System Health Science Center, 2121 West Holcombe Boulevard, Houston, TX 77030-3303. Phone: (713) 677-7522; Fax: (713) 677-7512; E-mail: wmckeeha{at}ibt.tamu.edu
6 The abbreviations used are: SMC, smooth muscle cell; FGF, fibroblast growth factor; FGFR, fibroblast growth factor receptor; RD, 1:1 RPMI and DMEM; FBS, fetal bovine serum; DT, Dunning tumor; DTE, DT tumor-derived epithelial cell; DTS, DT tumor-derived stromal cell; NPS, normal rat prostate; LDTS, low-passage DT tumor-derived stromal cell; HDTS, high-passage DT tumor-derived stromal cell; RT-PCR, reverse transcription-PCR; RPA, ribonuclease protection assay; nt, nucleotide.
7 C. Jin, F. Wang, X. Wu, C. Yu, Y. Luo, W. Chan, and W. L. McKeehan. Directionally-specific communication from prostate epithelium to stroma mediated by FGF9 and FGFR3, submitted for publication.
Received 10/21/02. Revised 6/ 4/03. Accepted 6/ 5/03.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-smooth muscle actin by transforming growth factor-ß 1 in quiescent human breast gland fibroblasts. Implications for myofibroblast generation in breast neoplasia. Lab. Investig., 68: 696-707, 1993.[Medline]
This article has been cited by other articles:
|
|
 |
|
  S. Oltean, B. S. Sorg, T. Albrecht, V. I. Bonano, R. M. Brazas, M. W. Dewhirst, and M. A. Garcia-Blanco Alternative inclusion of fibroblast growth factor receptor 2 exon IIIc in Dunning prostate tumors reveals unexpected epithelial mesenchymal plasticity PNAS, September 19, 2006; 103(38): 14116 - 14121. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Ohira, M. Sasaki, K. Harada, Y. Sato, Y. Zen, K. Isse, K. Kozaka, A. Ishikawa, K. Oda, Y. Nimura, et al. Possible Regulation of Migration of Intrahepatic Cholangiocarcinoma Cells by Interaction of CXCR4 Expressed in Carcinoma Cells with Tumor Necrosis Factor-{alpha} and Stromal-Derived Factor-1 Released in Stroma Am. J. Pathol., April 1, 2006; 168(4): 1155 - 1168. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Galie, C. Sorrentino, M. Montani, L. Micossi, E. Di Carlo, T. D'Antuono, L. Calderan, P. Marzola, D. Benati, F. Merigo, et al. Mammary carcinoma provides highly tumourigenic and invasive reactive stromal cells Carcinogenesis, November 1, 2005; 26(11): 1868 - 1878. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. A. Shain Exogenous Fibroblast Growth Factors Maintain Viability, Promote Proliferation, and Suppress GADD45{alpha} and GAS6 Transcript Content of Prostate Cancer Cells Genetically Modified to Lack Endogenous FGF-2 Mol. Cancer Res., November 1, 2004; 2(11): 653 - 661. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Jin, F. Wang, X. Wu, C. Yu, Y. Luo, and W. L. McKeehan Directionally Specific Paracrine Communication Mediated by Epithelial FGF9 to Stromal FGFR3 in Two-Compartment Premalignant Prostate Tumors Cancer Res., July 1, 2004; 64(13): 4555 - 4562. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Jin, K. McKeehan, W. Guo, S. Jauma, M. M. Ittmann, B. Foster, N. M. Greenberg, W. L. McKeehan, and F. Wang Cooperation between Ectopic FGFR1 and Depression |
, DTS2 in absence of FBS and FGF9; 
, DTS2 in absence of FBS plus FGF9; 
, DTS1 in the presence of 0.5% FBS and absence of FGF9; 
, DTS1 in the presence of 0.5% FBS and presence of FGF9 (
, DTS1 in the presence of 2% additional FBS and absence of FGF9; 
, DTS1 in the presence of both 2% FBS and FGF9. Response of DTS2 to FGF9 at days 3, 5, and 7 was significant at P < 0.01 relative to no FGF9. Response of DTS1 to 2% FBS was significant to P < 0.01, P < 0.05, and P < 0.01 at days 3, 5, and 7, respectively; bars, ±SD.