The signaling adapter protein PINCH is up-regulated in the stroma of common cancers, notably at invasive edges†
Article first published online: 5 SEP 2002
Copyright © 2002 American Cancer Society
Volume 95, Issue 6, pages 1387–1395, 15 September 2002
How to Cite
Wang-Rodriguez, J., Dreilinger, A. D., Alsharabi, G. M. and Rearden, A. (2002), The signaling adapter protein PINCH is up-regulated in the stroma of common cancers, notably at invasive edges. Cancer, 95: 1387–1395. doi: 10.1002/cncr.10878
Portions of this work were presented orally by Dr. Wang-Rodriguez at Experimental Biology 2000, San Diego, CA, April 17, 2000, and published in abstract form in FASEB J. 2000;14:A448.
- Issue published online: 5 SEP 2002
- Article first published online: 5 SEP 2002
- Manuscript Accepted: 19 APR 2002
- Manuscript Revised: 5 MAR 2002
- Manuscript Received: 13 NOV 2001
- Howard Hughes Medical Institute. Grant Number: 76296-54
- Cancer Research Coordinating Committee of the University of California
- Academic Senate, University of California, San Diego
- tumor-associated stroma;
- adapter protein;
- LIM domain;
PINCH is an LIM (double zinc finger domain) protein that functions as an adapter at a key convergence point for integrin and growth factor signal transduction. Because no information is available regarding its expression in vivo in human tissues, this study evaluated the distribution and abundance of PINCH in patients with breast, prostate, lung, colon, and skin carcinomas.
A polyclonal antibody was raised to a purified 6-histidine PINCH fusion protein and used to evaluate 74 cases comprising 33 breast carcinomas (21 ductal carcinomas, 6 lobular carcinomas, 4 ductal carcinomas in situ, 2 lobular carcinomas in situ), 22 prostate carcinomas, 5 colon carcinomas, 6 lung carcinomas (3 adenocarcinomas and 3 squamous carcinomas), and 8 skin carcinomas (4 basal cell carcinomas and 4 squamous cell carcinomas) by immunoperoxidase histochemistry of formalin-fixed, paraffin-embedded tissues. Lysates of frozen tissue from the epithelium of two normal breasts and six breast carcinomas were evaluated by immunoblotting.
Immunostaining for PINCH was increased in the cytoplasm of fibroblastoid cells in areas of the tumor-associated stroma in all carcinomatous tissues evaluated. The most intense stromal immunostaining for PINCH was noted at invasive edges, particularly in breast carcinomatous tissue. Immunoblotting of lysates from normal breast and breast carcinomatous tissue confirmed that PINCH protein expression was markedly increased in breast carcinomatous tissues.
PINCH is up-regulated in tumor-associated stromal cells, particularly at invasive edges, and may be a marker for stroma manifesting the ability to facilitate invasion. Because of this and because PINCH functions as a “molecular switch” in signal transduction, PINCH may be a new target for drug discovery aimed at the tumor-associated stroma. Cancer 2002;95:1387–95. © 2002 American Cancer Society.
Signaling complexes containing protein kinases, phosphatases, and their substrates are assembled within cells by noncatalytic adapter proteins.1 Adapter proteins orient molecules within signaling complexes to enhance the specificity or affinity of the enzymatic reactions and to localize proteins to specific compartments where there is a greater likelihood of interaction with downstream effectors. Adapters are not passive scaffolds for signaling complexes but rather have active roles in signal transduction and can act as oncogenes2–5 or tumor suppressors.6 Some adapter proteins are up-regulated in patients with specific cancers. For example, TRAF-1 is overexpressed in patients with chronic lymphocytic leukemia and with non-Hodgkin lymphomas7 and SETA is overexpressed in patients with astrocytomas, oligodendrogliomas, and oligoastrocytomas.8
Adapters function by virtue of protein-interacting domains. One such protein-interacting domain is the double zinc finger that characterizes the LIM family,9–11 so named because it was described originally in the Lin-11, Isl-1, and Mec-3 proteins.12, 13 One member of the LIM family, PINCH, consists primarily of five LIM domains,14 the largest number of LIM domains in a single protein.
Two proteins are associated directly with PINCH. The integrin-linked kinase (ILK),15, 16 a serine/threonine kinase involved in integrin signaling17, associates with the first PINCH LIM domain. The second protein, Nck-2, a Src homology adapter protein involved in growth factor signaling,18 associates with the fourth PINCH LIM domain. By virtue of these associations with ILK and Nck-2, PINCH participates in integrin and growth factor signal transduction at a key convergence point of these pathways.18 The proteins associating with the other eight zinc fingers in PINCH have not been identified yet. However, these proteins likely include downstream effectors of integrin and growth factor signaling.
Genetic studies in Caenorhabditis elegans have shown that loss of PINCH function results in a phenotype identical to that of integrin-null mutants,19 indicating that PINCH is required for integrin signaling to occur. PINCH is also required for ILK localization to integrin-containing adherens junctions20 where ILK regulates fibronectin matrix assembly.21 PINCH participates in growth factor signaling pathways via the direct associations of Nck-2 with activated forms of platelet-derived growth factor (PDGF) receptor-β, epidermal growth factor (EGF) receptor, and insulin receptor substrate-1.18
PINCH mRNA is expressed in most normal tissues14 and PINCH protein is localized to the cytoplasm and cell-matrix adherens junctions in cell lines.15 However, because no information is available regarding PINCH protein expression in vivo in tissues, this study evaluated the distribution and abundance of PINCH protein in a variety of common human cancers.
MATERIALS AND METHODS
Recombinant PINCH Fusion Proteins
A polymerase chain reaction (PCR) product corresponding to the entire open reading frame of human PINCH cDNA was ligated in-frame with the pAcSG His NT baculovirus transfer vector (PharMingen, San Diego, CA), which was then cotransfected with modified baculovirus DNA (Baculogold; PharMingen) into insect cells. Recombinant 6-histidine PINCH was purified by chromatography of an insect cell lysate on a metal chelate matrix (Probond; Invitrogen, La Jolla, CA). This protein was used as the immunogen to produce rabbit anti-PINCH.
A PCR product corresponding to the third PINCH LIM domain (residues 124–191) was directionally ligated with the pGEX-2TK plasmid vector (Pharmacia, Gaithersburg, MD) and used to transform competent Escherichia coli. The resulting glutathione S-transferase (GST) fusion protein (GST-PINCHLIM3) was purified by chromatography of E. coli lysate on glutathione Sepharose 4B (Pharmacia). This protein was used to affinity purify rabbit anti-PINCH to confirm its specificity.
A full-length PINCH maltose-binding protein fusion protein (MBP-PINCH)15 was provided generously by Dr. Chuanyue Wu (University of Pittsburgh, Pittsburgh, PA). This protein was also used to confirm the specificity of rabbit anti-PINCH.
Antibody to PINCH
Purified 6-histidine PINCH was subjected to sodium dodecyl sulfate (SDS)–polyacrylamide gel electrophoresis. The 38-kilodalton (kD) band corresponding to its calculated molecular weight was eluted and used as the immunogen. Rabbit antiserum to PINCH was produced at Rockland Laboratories (Gilbertsville, PA). Anti-PINCH IgG was purified from serum using Protein A-Sepharose chromatography (Pharmacia). For some experiments, anti-PINCH IgG was purified further by affinity chromatography on a GST-PINCHLIM3 fusion protein.
Tissues and Cell Lines
Excess human pathologic material was obtained from the clinical laboratories of the University of California, San Diego Medical Center, and the Veteran's Administration Medical Center (La Jolla, CA) according to guidelines established by the respective institutional review committees.
The material examined consisted of formalin-fixed, paraffin-embedded tissues from 74 cases. There were 33 breast carcinoma cases including 21 ductal carcinomas, 6 lobular carcinomas, 4 ductal carcinomas in situ (DCIS), and 2 lobular carcinomas in situ (LCIS). In addition, there were 22 prostate carcinomas, 5 colon carcinomas, 6 lung carcinomas (3 adenocarcinomas and 3 squamous carcinomas), and 8 skin carcinomas (4 basal cell carcinomas and 4 squamous cell carcinomas). More than one tissue block per case was tested (when available) so that more than 100 blocks were examined and more than one section was examined from each block so that more than 200 sections were tested.
Also examined were surgical tissues frozen at −80 °C (samples included one muscle biopsy specimen, two reduction mammoplasty specimens, and six breast carcinoma specimens). Cultured cells were obtained from Dr. Mark Kamps (human breast cell line HBL-100 and mouse fibroblast cell line NIH3T3), Dr. Colin Bloor (rat cardiac myoblast cell line H9c2 and cultured pig aortic smooth muscle cells), and Dr. Paul Stein (human prostate cancer cell lines LNCaP and PC-3).
Tissues and cultured cells were solubilized in 1% SDS/phosphate-buffered saline (PBS) or in 1% Triton X-100/PBS, both containing a protease inhibitor cocktail (Complete; Roche, Indianapolis, IN), or in 150 mM NaCl, 50 mM HEPES, pH 7.5, containing 1% IGEPAL CA-630, 0.5% sodium deoxycholate, 0.1 mg/mL phenylmethylsulfonyl fluoride, 1 μg/mL aprotinin, and 1 μg/mL leupeptin. Protein concentrations were determined by the DC protein assay (BioRad, Richmond, CA). Solubilized proteins were separated by electrophoresis in 10% SDS-polyacrylamide gels and transferred to nitrocellulose (Hybond-ECL; Amersham, Arlington Heights, IL) by electroblotting in 25 mM Tris, 192 mM glycine, and 20% methanol, pH 8.3.
Nitrocellulose membranes were blocked for 30 minutes with 5% nonfat dried milk in Tris-buffered saline (TBS) containing 0.1% Tween 20 (TBS-T), pH 7.5, and reacted overnight at 4 °C with rabbit anti-PINCH at 1 μg/mL in 5% nonfat milk/TBS-T. Reactions were detected using horseradish peroxidase (HRP)-conjugated anti-rabbit Ig (Amersham) diluted at 1:3000 for 30 minutes at room temperature (RT) followed by enhanced chemiluminescence (ECL) (Amersham). Binding of mouse anti-polyhistidine clone HIS-1 (Sigma, St. Louis, MO) was detected using HRP-conjugated anti-mouse Ig (Amersham) diluted at 1:1000 for 30 minutes. Biotinylated molecular weight markers (Amersham) were detected using an HRP-streptavidin complex (Amersham).
Deparaffinized and rehydrated tissue sections were reacted with 3% hydrogen peroxide for 5 minutes to block endogenous peroxidase and with an antigen retrieval solution (Dako, Carpinteria, CA) for 5 minutes at 95 °C. This was followed by an additional reaction for 20 minutes at RT. Sections were washed in TBS, blocked with 1.5% horse serum/TBS, and reacted for 1 hour at RT with rabbit anti-PINCH IgG in 1.5% horse serum/TBS. Reactions were developed using the Universal Vectastain Elite ABC kit and 3,3′-diaminobenzidine (both from Vector, Burlingame, CA). Staining was positive if it exceeded that on a control section from the same block reacted with nonimmune rabbit IgG (Vector) used at the same concentration as the primary antibody (1 μg/mL).
Staining of the purified HIS-PINCH fusion protein with anti-HIS (Fig. 1, lane 1) showed the expected 38-kD band as well as a higher molecular weight band around 75 kD. The same pattern was noted in insect cell lysates containing HIS-PINCH immunostained with anti-PINCH (lane 2). One explanation for this is that HIS-PINCH migrated in the polyacrylamide gel as a monomer and dimer. Dimerization, if it occurred, was not the result of inadequate reduction as samples were boiled for 5 minutes in 2-mercaptoethanol and dithiothreitol before loading. However, it is possible that HIS-PINCH self-associated even under reducing conditions, as do members of the glycophorin family.22 In addition to detecting the HIS-PINCH fusion protein used as its immunogen, anti-PINCH also detected MBP-PINCH (lane 3) and GST-PINCHLIM3 (data not shown).
Anti-PINCH detected a band corresponding to the molecular weight expected for native PINCH (37 kD) in cultured cell lysates (Fig. 1, lanes 4–8). It also detected a 37-kD band in a lysate prepared from a frozen human skeletal muscle biopsy specimen (Fig. 1, lane 9). The intensity of immunostaining was evidence that PINCH was abundant in skeletal muscle, which is consistent with the high PINCH mRNA expression detected in skeletal muscle by Northern analysis.14
As has been found with other PINCH antibodies,15 the anti-PINCH produced for this study sometimes detected a higher molecular weight band at about 75 kD (Fig. 1, lanes 4,5,7). This occurred even though the immunogen for this antibody was the 38-kD band of purified HIS-PINCH fusion protein eluted from the polyacrylamide gel. It is likely that the 75-kD band in cultured cell lysates corresponds to the dimer of native PINCH, just as the same sized band in transfected insect cell lysates corresponds to the HIS-PINCH dimer (Fig. 1, lanes 1,2).
Anti-PINCH IgG immunoaffinity purified on GST-PINCHLIM3 detected no PINCH in lysates from two normal breast tissue samples. However, abundant PINCH was detected in lysates from six breast cancer tissue samples (compare lanes 10 and 11 in Fig. 1), consistent with increased PINCH protein expression in breast cancer tissue (compare immunohistochemical staining for PINCH in normal breast tissue [minimal staining, see Fig. 3G] and in breast cancer tissue [marked stromal staining, see Fig. 3F]). Anti-PINCH IgG immunoaffinity purified on GST-PINCHLIM3 (Fig. 1, lane 11) showed the same immunoblot pattern as anti-PINCH IgG (Fig. 1, lanes 4–9), supporting the PINCH specificity of the antibody.
In the tissues examined (with the exception of skin), normal epithelial and carcinomatous cells showed minimal PINCH immunostaining (Figs. 2, 3). Normal skin tissue showed moderate immunostaining for PINCH in the epidermis, hair follicle sheaths, and scattered dermal fibroblasts, but no immunostaining of collagen or other matrix components (Fig. 2E). The distribution of PINCH protein in skin as shown by immunohistochemistry (this study) is identical to that reported for PINCH mRNA in skin by in situ hybridization,23 supporting the tissue specificity of the PINCH antibody used in this study. Basal cell carcinoma and squamous cell carcinoma also showed moderate PINCH immunostaining (Fig. 2F).
Immunostaining for PINCH was minimal in normal tissue stroma but increased in areas of the tumor-associated stroma in all carcinomatous tissue samples examined (Figs. 2,3). PINCH immunostaining was especially intense in stroma at invasive edges, particularly in breast carcinoma tissue samples (Fig. 3). Examination of higher power views showed that most of the PINCH immunostaining was contained within the cytoplasm of cells that appeared to be fibroblasts (Fig. 2D).
Comparison of the immunohistochemistry and immunoblotting results for breast tissue is especially instructive. Normal breast tissues showed minimal staining for PINCH by both techniques and therefore the expression of PINCH protein in normal breast tissue samples is low. Breast carcinoma tissue samples, however, showed significant up-regulation of PINCH protein by immunoblotting (Fig. 1, lane 11). The distribution of the increased PINCH protein was found primarily in the stroma associated with breast carcinoma and most notably in the stroma directly adjacent to the invasive edge. Increased PINCH in the stroma at invasive edges occurred in cases where the breast carcinoma cells invaded as a cohort (see the ductal breast carcinoma section in Fig. 3B) and as individual cells (see the lobular breast cancer section in Fig. 3D).
PINCH immunostaining of the stroma adjacent to noninvasive carcinoma in situ lesions was variable (Fig. 4). Some DCIS lesions were encircled by a band of PINCH-stained stroma (Fig. 4A) whereas other DCIS lesions showed minimal stromal immunostaining. LCIS (Fig. 4B) and prostatic intraepithelial neoplasia (Fig. 4C) showed variable but generally modest stromal immunostaining for PINCH. PINCH immunostaining was abundant in the stromal cells of a healing wound (Fig. 4D).
This study shows that the LIM family adapter protein PINCH is up-regulated in the stroma associated with common carcinomas. Immunostaining for PINCH was particularly intense in stromal cells at invasive edges, suggesting that the up-regulation of PINCH may be a marker for stroma that can facilitate cancer invasion.
The tumor-associated stroma consists of fibroblasts, myofibroblasts, smooth muscle cells, vascular elements, inflammatory cells, and extracellular matrix (ECM) and is important in facilitating cancer growth and invasion.24–30 Some markers in the tumor-associated stroma have prognostic value, including oncofetal fibronectin,31 basic fibroblast growth factor,32 c-MET/hepatocyte growth factor receptor,33 fibroblast activation protein/seprase,34 hyaluronan,35, 36 syndecan-1,37 chondroitin sulphate,38 and galectin-1.39 It has not been determined yet if increased PINCH immunostaining is a clinically relevant prognostic marker. However, the intense PINCH stromal immunostaining noted at invasive edges suggests that this could be the case.
Tissue remodeling in cancer and wound healing involves increased deposition as well as increased proteolysis of ECM, leading to the observation that cancer is a “wound that does not heal.”24 Assembly of a major ECM component, fibronectin, is regulated by ILK21 in a process that requires PINCH for its localization to the integrin-containing adherens junctions.20 Increased PINCH protein expression in the tumor-associated stroma may therefore be associated with enhanced fibronectin matrix assembly there. The fibronectin matrix and other ECM components deposited by stromal cells migrating in advance of the cancer leading edge provide the appropriate surface on which the carcinomatous cells can migrate.40, 41 The most intense immunostaining for PINCH was noted in stromal cells just in front of and at the leading edge of cancer invasion (Fig. 3), which is consistent with this concept.
Growth factors also play an important role in tumor-stromal interactions. For instance, PDGF-mediated tumor-stromal interactions are well documented in patients with breast, lung, and skin carcinoma in whom these interactions are viewed as essential to tumor growth.42–45 PINCH is involved in PDGF signaling (as well as in signaling by other growth factors such as EGF and insulin) by virtue of its interaction with Nck-2.18 Therefore, increased PINCH protein expression may be associated with up-regulated growth factor signaling in stromal cells.
How could up-regulation of an adapter protein such as PINCH lead to increased signaling in the tumor-associated stroma? A hypothesis proposed by Kholodenko et al.46 regarding the role of adapter proteins in signal transduction provides a possible explanation. In their view, the major mechanism underlying increased signaling is an increase in the number of signaling complexes. They proposed that membrane anchoring of a single member of a signaling complex fails to increase the number of complexes, whereas membrane anchoring of two members of a signaling complex leads to a marked decrease in the dissociation constant and a significant increase in the number of complexes. The required membrane anchoring can be accomplished by the binding of an adapter protein to one membrane-anchored component followed by binding of the same adapter protein to a second membrane-anchored component, a process they called “piggyback riding.”
PINCH may be a piggyback riding adapter protein, associating first with integrin receptors via ILK and subsequently with activated growth factor receptors via Nck-2 (or vice versa). The association of two membrane-anchored proteins (integrin receptor and activated growth factor receptor) would lead to the production of more PINCH-containing signaling complexes than would the association of PINCH with either receptor alone. These stable PINCH-containing complexes could then recruit downstream effectors, some of which would likely associate with PINCH LIM domains.
This model of PINCH function is supported by the known direct associations of the proteins involved15, 17, 18, 20 and the colocalization of integrins with activated growth factor receptors.47 PINCH is envisioned as a “rate-limiting” component to signal transduction. Under normal conditions of low PINCH abundance, there is a low level of signaling, but when the PINCH protein is increased, augmented signaling occurs. It would not be necessary for the protein levels of other components in the PINCH signaling complex to increase for there to be an increase in the amount of signaling. A change in the amount of PINCH alone would be sufficient because it would permit the organization of existing components into functional signaling complexes.
The results reported in this study suggest that PINCH is present in normal stromal cells in limiting amounts. However, when up-regulated in tumor-associated stromal cells, PINCH may participate in a signaling complex that brings together integrin and growth factor signaling pathways and recruits their downstream effectors. If PINCH up-regulation could be prevented or the protein itself inhibited, then integrin and growth factor signaling pathways might be down-regulated simultaneously. Therefore, PINCH may represent a new target for drug discovery aimed at the tumor-associated stroma.
The authors gratefully acknowledge the technical assistance of Catherine Sincich and Andrea Pakula.
- 34Stromal expression of fibroblast activation protein/seprase, a cell membrane serine proteinase and gelatinase, is associated with longer survival in patients with invasive ductal carcinoma of breast. Int J Cancer. 2001; 95: 67–72., , , , .
- 40Intercellular invasion and the organizational stability of tissues: a role for fibronectin. Biochim Biophys Acta. 2000; 1470: 9–20., .