Abstract
The neuregulin 4 gene encodes at least five different variants (designated A1, A2, B1, B2 and B3) produced as a result of alternative splicing. We have determined their sites of expression in normal human adult tissues using isoform-specific antibodies. Their expression is cell type specific and differs in subcellular location suggesting that they may have varied functions in these contexts. We have shown in a panel of prostate cancers that each form is present to differing degrees, and that principal component analysis indicates that there are three patterns of expression. Some isoforms were positively correlated with high prostate-specific antigen levels and others were inversely associated with Gleason score. Synthetic, refolded A forms promoted lamellipodia and filopodia formation in cells expressing the ErbB4 (CTa) receptor and stimulated cell motility in wound healing assays. The data suggest that the different forms have varied sites of expression and function, and this includes effects on cell architecture and motility.
Introduction
The four ErbB growth factor receptors, ErbB1 (HER1), ErbB2 (HER2), ErbB3 (HER3) and ErbB4 (HER4), are tyrosine kinases that are predominately expressed in epithelial and mesenchymal cells where they regulate a host of normal processes, which include cell division, differentiation, apoptosis and motility (Yarden & Sliwkowski 2001, Citri & Yarden 2006).
To date, eleven ligands for these receptors have been identified which can be grouped into three classes, those that only interact with ErbB1, those that interact with ErbB1 and ErbB4 and those that interact with ErbB3 or ErbB4. The ligands in the latter class are the neuregulins (Jones et al. 1999, Hynes et al. 2001), which are encoded by four genes (NRG1–NRG4; Falls 2003, Stein & Staros 2006). These undergo extensive alternative splicing, permitting the expression of proteins which have diverse motifs, subcellular localisations and functions (Hayes & Gullick 2008). The NRG1 gene can express more than 15 products under the control of nine alternative promoters and six sites of initiation of transcription (Steinthorsdottir et al. 2004). The NRG2 and NRG3 gene products have been less well studied, but at least six alternative splice variants have been described for NRG2 and two for NRG3 (Hayes & Gullick 2008). Recently, five gene products have been described for the NRG4 gene, which all share the same translation initiation site (Hayes et al. 2007). Two of these variants, NRG4A1 and NRG4A2, have a transmembrane region and encode a shared biologically active 8 kDa fragment that includes an epidermal growth factor (EGF) motif (Hayes et al. 2008). The C-terminal sequence of NRG4A2 has a predicted PDZ (PSD-95/Discs-Large/ZO-1)-binding motif, but as yet the binding partner (if any) has not been identified. The other three NRG4 variants (NRG4B1–3) all share a common truncated EGF domain that only includes the first four cysteine residues in the EGF motif and lacks the transmembrane region.
NRG1 can bind to ErbB3 and, by promoting heterodimerisation with ErbB2, stimulate signalling via the PI3 kinase/Akt pathway (Kita et al. 1994, Citri et al. 2003). A chemically synthesised and refolded EGF domain of NRG4 binds directly to ErbB4 and stimulates its phosphorylation (Harari et al. 1999, Hayes et al. 2007). Four human ErbB4 isoforms have been identified, two of which permit proteolytic cleavage and translocation to the nucleus of the subsequently released cytoplasmic domain (Carpenter & Liao 2009). The other alternative products either contain an additional 16 amino acids (CTa) in the intracellular domain that includes the sequence YTPM, a consensus motif for a PI3 kinase binding site or do not (CTb) (Elenius et al. 1999, Jones 2008, Sundvall et al. 2008). Activation of both the latter forms promotes cell proliferation, but only the ErbB4CTa induces the reorganisation of actin filaments into membrane ruffles and lamellipodia, a process required for cell migration (Kainulainen et al. 2000).
Numerous studies have implicated the aberrant behaviour of the ErbB receptors and the neuregulins in several human diseases including epithelial cancers (Montero et al. 2008). It is therefore important that the components of this exquisitely intricate system and the subcellular localisation of these isoforms are identified to enhance understanding of the NRG4 family of ligands and to identify novel molecular targets to facilitate targeted treatment of cancer patients. We have shown in previous work using specific antibodies that the NRG4A isoforms are synthesised in prostate cancer cells (Hayes et al. 2007). There have, however, been no reports describing the distribution of the recently identified NRG4 isoforms in normal human tissues or their relative levels of expression in cancer. We show here that NRG4A1, NRG4A2 and NRG4B3 are all expressed in prostate cancer cells, and that the subcellular distribution of these isoforms is different. A principle component analysis indicates that the cancers fall broadly into three patterns based on their expression or lack of expression of the various isoforms.
We have also demonstrated that high levels of expression are associated with advanced tumour stage and tumour-node metastasis (TNM) status (Hayes et al. 2007). In this work, we show a relationship between the expression of NRG4A1 and prostate-specific antigen (PSA) levels suggesting a potential role in tumour growth and metastasis. We therefore tested the hypothesis that NRG4 may play a role in cell motility by assessing the abilities of the EGF-like domains of NRG4A and NRG4B isoforms to promote membrane ruffling, actin reorganisation and cell migration.
Materials and methods
Clinical samples and immunohistochemistry
Two tissue microarrays from SuperBioChips (Yongdu-dong, Seoul, Korea) were used in this study. The first (AC1) was an array of 30 pairs of different normal human tissues taken from the organs of 60 non-cancer patients. The second (CA3) contained 40 cases of human prostate adenocarcinoma, and in nine of the cases, matched tissues were taken from adjacent normal prostate. The Gleason grade was determined by M M B and G A R. Ethical approval for their use was obtained from the research ethics committee at the University of Kent.
Antibodies and scoring
Five rabbit NRG4 polyclonal antibodies were made to synthetic peptides (Hayes et al. 2007). The anti-127 antibody recognised the homologous N-terminal of all the NRG4 isotypes, anti-123 recognised the C-terminal of the EGF-like domain and is NRG4A1 and NRG4A2 specific, anti-128 recognised the unique C-terminal of NRG4A1, anti-135 recognised the unique C-terminal of NRG4A2 and anti-134 recognised the unique C-terminal of the NRG4B3 isoform (Fig. 1). The HFR-1 antibody was raised to the intracellular sequence (aa1249–1264) of ErbB4 (Srinivasan et al. 1998). An antibody that recognised the ErbB4CTa splice variant was made to the peptide (113.4L -SEIGHSPPPAYTPMSG), which is only present in the CTa variant. An anti-RhoA mouse monoclonal antibody (STA-403-A) was obtained from Cell Biolabs, Inc. (San Diego, CA, USA).
Schematic representation of NRG4 isoforms and position of peptides used to raise antibodies. The A isoforms of NRG4 have an intact EGF-like domain and transmembrane domain (TM). NRG4B3 has a partial EGF-like domain. The different C-terminal sequences are shown. The regions used for immunisation are underlined, and the numbers represent antibody designation.
Citation: Endocrine-Related Cancer 18, 1; 10.1677/ERC-10-0112
Staining was performed using peptide antibodies with the StreptABC complex/Duet kit (Dako, Ely, Cambridgeshire, UK). Scoring was done independently by M M B and G A R, and any discrepancies were resolved by discussion. Scoring was performed as described by Rajkumar et al. (1996). A score of 0–3 was given for intensity of staining, and a score of 1–4 was used for indicating the percentage of cells positive by quartiles. The data consist of measurements from 49 samples. There are samples of cancer tissue only from 31 patients and a sample of cancer tissue and matched normal tissue from a further nine patients. Estimates of the extent and intensity of staining from the five peptide-specific antibodies as described above were collected for each sample. Data were missing for sample 22 for anti-128, anti-134 and anti-135 and for sample 12 for anti-135 as only fragments of tissue that were too small to evaluate were present.
Statistical analysis
The degree of heterogeneity of expression of these molecules in prostate cancer is not known. We therefore chose to score the percentage of cells, which were positive, as well as the level of expression. We analysed these separately to determine whether each measurement could be related to the selected co-variables. Associations between the extent and intensity of staining with covariates were assessed by Spearman's rank correlation coefficient, and as multiple comparisons were made, a conservative P value was used. Principal component analysis of the Spearman's rank correlation matrix for the scores of each neuregulin antibody was used to investigate whether the expression of a given neuregulin was related to the expression of any other. The samples were objectively allocated to clusters by their immunoreactivity scores using a single-link clustering algorithm, based on standardised Euclidean distances.
Tissue culture
NIH3T3ErbB4CTa cells (a mouse embryonic fibroblast cell line stably transfected with ∼106 ErbB4 receptors per cell) (Cohen et al. 1996) and Cos-7 cells (African green monkey fibroblast-like cells) were cultured in DMEM (Gibco), 10% (v/v) FCS (Gibco), 2 mM l-glutamine, 50 μg/ml penicillin and 50 μg/ml streptomycin. The cells were fed every 3–4 days and subcloned every 7 days.
Transfections
For SDS-PAGE, Cos-7 cells were seeded in 3.5 cm dishes to reach ∼60% confluency after 17–24 h. Cells were transfected with 1 μg DNA (pcDNA 3.1ErbB4CTa, pcDNA3.1ErbB4CTb or pcDNA3.1 only) using the liposomal transfection reagent FuGENE 6 (Roche Molecular Biochemicals).
Immunofluorescence
Cells were washed in PBS, fixed with 4% (w/v) paraformaldehyde/PBS for 10 min and permeabilised with 0.1% (v/v) Triton/PBS for 5 min at 4 °C. Actin was detected using Phalloidin Alexa Fluor 546 (Molecular Probes, Paisley, Renfrewshire, UK). Images were obtained using Leitz DMRB immunofluorescence microscope and ×10 and ×40 Fluotar lenses and an HCX PL Apo ×63 lens.
Sample preparation, electrophoresis and immunoblotting
NIH3T3 cells (transfected as described above) were washed twice with PBS and then directly solubilised in 400 μl 2× Laemmli sample buffer, boiled for 5 min, and SDS-PAGE was carried out followed by immunoblotting onto nitrocellulose (Amersham).
Synthesis and peptide refolding
Four NRG4-specific peptides (shown in Table 1) were synthesised by the Protein Science Unit, Research School of Biosciences, University of Kent and refolded as described previously (Hayes et al. 2007).
Four NRG4-specific peptides
| NRG4 variant | Peptide sequence |
|---|---|
| NRG4A1/A2 | DHEQPCGPRHRSFCLNGGICYVIPTIPSPFRCCIENYTGARCEEVFL |
| NRG4B1 | DHEEPCGPSHKSFCLNGGLCYVIPTIPSPFCRK |
| NRG4B2 | DHEEPCGPSHKSFCLNGGLCYVIPTIPSPFCS |
| NRG4B3 | DHEEPCGPSHKSFCLNGGLCYVIPTIPSPFCSLHENENDNNEDLYDDLLPLNE |
Wound migration assay
Growth factor-stimulated cell migration was monitored using the CellPlayer Cell Migration 96-well assay, which was run using the automated Incucyte FLR live cell imaging system (Essen Bioscience, Welwyn Garden City, Hertfordshire, UK) NIH3T3ErbB4CTa cells were seeded at ∼3.5×103/well (in low-serum media – DMEM (Gibco), 0.1% (v/v) FCS (Gibco), 2 mM l-glutamine, 50 μg/ml penicillin and 50 μg/ml streptomycin) in an Essen Imagelock 96-well plate and were allowed to reach confluency (24 h). A wound was made in the monolayer of cells using a 96-well wound maker tool with PTFE pin tips according to the manufacturer's protocol. The media were removed, and the cells were washed twice in Hanks' balanced salt solution (HBSS, Gibco) to remove dislodged cells. To the wounded monolayer, 1×10−6 M NRG4 isotype-specific peptide (in low-serum media – DMEM as described above) was added, and the 96-well plate was inserted into the IncuCyte FLR platform (which includes an automated microscope) and incubated in 5% CO2 at 37 °C. High-density phase contrast images were automatically captured using a ×20 lens at 1 h time intervals, and the data were analysed using the IncuCyte software. The metric, relative wound density (RWD), which measures the relative cell density between the wounded area and the non-wounded at each time point, was used to assess the rate of cell migration.
Results
Distribution of NRG4 isoforms in normal tissues
An analysis of the expression of different NRG4 isoforms was performed using a panel of NRG4 antibodies (Fig. 1) in 30 different normal tissues taken from organs from 60 individual non-cancer patients. These results show that they frequently have distinct subcellular localisations and are often cell specific, which suggests that different isoforms will have different functions, e.g. in the skin, the epidermal positivity was basal with anti-128 and was diffuse with anti-135 with a greater intensity of staining in the upper half; in contrast, dermal fibroblast positivity was limited to the upper third with anti-128 and was diffuse with anti-135 (Fig. 2, panel A). Of particular note is the expression of NRG4B3 in the endothelial cells of a wide range of tissues (Supplementary Figure 1 (5–7), see section on supplementary data given at the end of this article) and nuclear staining with anti-123 in the salivary gland (data not shown). These results are summarised in Table 2, and more examples of immunohistochemical staining and a detailed description of their expression in a range of tissues are given in the Supplementary text and Supplementary Figure 1, see section on supplementary data given at the end of this article.
(Panel A) Immunohistochemical detection of the NRG4A isotypes in normal skin tissue. In the epidermis, anti-128 (NRG4A1) immunoreactivity was most intense in the basal cells, arrow (b), and anti-135 (NRG4A2) immunoreactivity was diffuse, with a greater intensity of staining in the upper half (arrow). In the dermal layer, anti-128 fibroblast positivity was confined to the upper third and diffuse with anti-135. (Panel B) Distinct subcellular distribution of NRG4A isotypes observed in the prostatic luminal epithelium. Immunoreactivity of the anti-123 antibody (to a homologous region of NRG4A1 and NRG4A2) is either granular, membranous or diffuse from three individual cases of prostate cancer. (Panel C) Stimulation of NIH3T3ErbB4CTa cells by NRG4A1/A2 peptide promotes membrane ruffling. NRG4A1/A2 peptide was added to NIH3T3ErbB4CTa cells, and the redistribution of actin (red) and RhoA (green) was determined by immunofluorescence and compared to untreated cells. Membrane ruffling can be observed in cells treated with NRG4A1/A2 and NRG1β1 (arrows); however, cells treated with NRG4B1 or unstimulated cells do not show membrane ruffling.
Citation: Endocrine-Related Cancer 18, 1; 10.1677/ERC-10-0112
Distribution of NRG4 isoforms in normal human tissues
| Tissue | Anti-123 NRG4A1/A2 | Anti-127 pan | Anti-128 NRG4A1 | Anti-134 NRG4B3 | Anti-135 NRG4A2 |
|---|---|---|---|---|---|
| Skin | |||||
| Epidermis | − | − | ++ | − | + |
| Digestive system | |||||
| Salivary gland | |||||
| Ducts | ++ | + | ++ | + | + |
| Acini | − | − | − | − | + |
| Oesophagus | − | − | − | − | − |
| Stomach | ++ | + | +++ | + | + |
| Small intestine | ++ | + | + | + | − |
| Colon | + | − | − | − | − |
| Liver | − | − | − | + | − |
| Gallbladder | ++ | − | + | + | + |
| Pancreas | |||||
| Acini | ++ | ++ | ++ | − | ++ |
| Islets | − | − | − | ++ | − |
| Cardiovascular system | |||||
| Heart myocytes | + | − | − | − | − |
| Aorta muscle cells | + | + | + | + | + |
| Lymphoreticular system | |||||
| Lymph nodes | |||||
| Macrophages | + | − | + | ++ | + |
| Plasma cells | − | − | ++ | − | − |
| Tonsil lymphoid cells | + | − | ++ | − | − |
| Thymus adipocytes | + | ++ | + | + | − |
| Spleen lymphoid cells | ++ | − | − | − | − |
| Respiratory system | |||||
| Lung alveolar macrophages | + | ++ | ++ | ++ | + |
| Bronchiolar epithelium | ++ | − | − | − | − |
| Pneumocytes | − | + | + | − | − |
| Nervous system | |||||
| Cerebral cortex | − | − | − | − | − |
| Cerebellum | − | − | − | − | − |
| Urogenital system | |||||
| Kidney glomeruli | − | − | − | − | − |
| Proximal tubules | + | − | − | − | − |
| Distal tubules | − | + | + | − | − |
| Prostate | + | − | + | + | − |
| Uterus (endometrium) | − | − | − | − | − |
| Placenta (trophoblastic epithelium) | ++ | + | + | + | − |
| Breast (lobular epithelium) | + | − | − | − | − |
| Umbilical cord | − | − | − | + | − |
−, Negative; +, weak staining; ++, moderate staining; +++, strong staining.
Distribution of NRG4 isoforms in prostate cancer
We have shown previously on a single case of prostate cancer that the panel of antibodies were able to detect specifically the splice variants (Hayes et al. 2007). Here, we use a panel of 40 cases of prostate cancer and matched normal tissues from nine of the cases to determine the levels and proportion of cancers cells expressing the different splice variants of NRG4 (254 measurements). In most positive tumour cells, the staining was diffuse and homogenous in the cytoplasm (Supplementary Figure 2, see section on supplementary data given at the end of this article), but in some it was most intense at the luminal surface (data not shown). However, in different patients, three different subcellular distributions were observed with the anti-123 antibody granular, membranous and diffuse (Fig. 2, panel B). Except for the 123-antibody, <20% of tumours showed moderate and strong positivity with the other antibodies (Table 3). Strong staining was observed only in a small number of tumours with antibodies to anti-123 and anti-128. There was patchy, weak and moderate staining of stromal cells and smooth muscle in the walls of blood vessels with all antibodies. Nerve fibres stained weak to moderately in five tumours, and there was moderate and strong staining of endothelial cells in blood vessels in all tumours with anti-134. Almost 70% of the cases were negative for anti-135 (NRG4B3). A comparison of the nine cases, where matched normal tissue and tumour (from the same patient) were analysed, showed that in some cases the staining in the tumour was more intense with several of the antibodies and in others it was less intense. There was no anti-135 staining observed in the normal tissues. It was not possible with such small numbers to reveal any trends for altered expression, and a larger series would be required to explore this issue.
Distribution of the strength of NRG4 isoform-specific staining in human prostate cancer cases
| Intensity of staining | Anti-123 (%) | Anti-127 (%) | Anti-128 (%) | Anti-134 (%) | Anti-135 (%) |
|---|---|---|---|---|---|
| Negative | 7.5 | 55 | 46.2 | 53.8 | 68.4 |
| Weak | 40 | 45 | 38.5 | 46.2 | 23.7 |
| Moderate | 45 | 0 | 12.8 | 0 | 7.9 |
| Strong | 17.5 | 0 | 2.5 | 0 | 0 |
Statistical analysis
Using a 1% significance level, taken separately, the extent and intensity of staining were highly correlated (P<0.01) within each antibody (123, ρ=0.480; 128, ρ=0.896; 127, ρ=0.946; 134, ρ=0.945 and 135, ρ=0.975). A principal component analysis for each of the five antibodies was carried out on a Spearman's rank correlation table. This exploratory analysis suggests the possibility of three types of sample, one is distinguished by a large number of negative results, particularly for anti-135, anti-134 and anti-127, one is mainly positive for all the antibodies and the other is intermediate (Supplementary Figure 3A, see section on supplementary data given at the end of this article). It was also noted that anti-127 and anti-134 make similar contributions to the first two principal components. This suggests that intact NRG4B3 can be identified and is not cleaved, i.e. the antibodies that recognise the N-terminal (anti-127) and the C-terminal (anti-134) behave in a similar way. The strength and intensity of immunoreactivity for anti-123 (NRG4A1/A2) seem to sometimes behave similarly to anti-128 (NRG4A1) and sometimes to anti-135 (NRG4A2) (Supplementary Figure 3B, see section on supplementary data given at the end of this article). This suggests that both NRG4A1 and NRG4A2 are expressed in these cancers, but cleavage of the N-terminal has occurred as there is no correlation with anti-127 (the common N-terminal).
Owing to the relatively small number in the three groups, we analysed the complete set of cases for any associations with co-variables. There were three significant correlations with the covariates. The extent of staining for anti-128 was significantly positively correlated with PSA at rs (adjusted for ties)=0.35, P=0.05. The Gleason score was significantly negatively correlated with quantity (rs=−0.45, P=0.00) and extent (rs=−0.44, P=0.01) of staining for anti-127.
Exogenous refolded NRG4 peptide promotes membrane ruffling at the leading edge of NIH3T3ErbB4CTa cells
NRG1 stimulation of ErbB4CTa can activate PI3 kinase (Elenius et al. 1999), and NRG1 induces chemotaxis with a concomitant reorganisation of actin filaments into lamellipodia and membrane ruffles (Ritch et al. 2003). Also, in cancer cells that show an abnormal production of NRG1, there is often aberrant gene regulation which results in increased cell motility (Kainulainen et al. 2000). As we have previously observed that overexpression of NRG4A1 in Cos-7 cells promotes the formation of NRG4A1-enriched membrane ruffles (Hayes et al. 2008), it was of interest to elucidate whether exogenous addition of the novel NRG4 isoforms could promote membrane ruffling, actin reorganisation and cell migration in NIH3T3 cells stably transfected with ErbB4CTa.
The 113.4L ErbB4CTa-specific antibody confirmed that the NIH3T3 cells overexpressed only the CTa isoform of ErbB4 and not the CTb isoform (Fig. 3). Exogenous addition of NRG4A1/A2 to serum-starved NIH3T3ErbB4CTa promoted the formation of actin-rich areas at the leading edge of the migrating cell. This rearrangement of the cortical actin was not observed with the NRG4B isoforms, Fig. 2 panel C (data shown for NRG4B1 only) or in unstimulated cells. As it has previously been observed in growth factor-stimulated NIH3T3 cells that activated RhoA is localised to membrane ruffles (Kurokawa & Matsuda 2005), co-localisation of RhoA and actin confirmed the presence of membrane ruffles at the leading edge in NRG4A1/A2-stimulated NIH3T3ErbB4CTa cells (Fig. 2 panel C). In addition, remodelling of the actin cytoskeleton was observed in cells treated with NRG4B1, there was dissociation of the F-actin filaments and vesicles decorated with actin were distributed throughout the cytoplasm.
Identification of the ErbB4CTa isotype in stably transfected NIH3T3 cells. The ErbB4CTa-specific antibody (113.4L) identified a 180 kDa protein in stably transfected NIH3T3ErbB4 cells (lane 1). Transient transfection of Cos-7 cells with ErbB4CTa (lane 3) and ErbB4CTb (lane 4) confirms the specificity of the 113.4L antibody for the ErbB4CTa isotype. The anti-ErbB4 antibody HFR-1 detects both ErbB4CTa (lane 3) and ErbB4CTb (lane 4). The negative controls were transfected with empty pcDNA3.1 plasmid (lane 2).
Citation: Endocrine-Related Cancer 18, 1; 10.1677/ERC-10-0112
NRG4A1/A2 promotes ErbB4CTa-activated cell migration
The previous results show that NRG4A1/A2 stimulation of ErbB4CTa leads to the formation of membrane ruffles at the leading edge, which is a process involved in directed cell migration. A directed wound healing assay using NIH3T3ErbB4CTa cells was used to evaluate whether any of these NRG4 isoforms could promote cell migration in vitro. After 29-h incubation, cells that had been incubated with the NRG4A1/A2 peptide migrated sufficiently for wound closure to occur and had a RWD of 71% compared with 57% for untreated cells. Surprisingly, cells treated with the B isoforms of NRG4 showed retarded cell motility compared with the untreated cells. Cells treated with either the NRG4B1 or the NRG4B3 peptide had a RWD of 43 and 34% respectively (Fig. 4A) and showed no further migration throughout the time course. In contrast, by 45 h the RWD for NRG4A1/A2-stimulated cells had increased to 92% compared with 76% for untreated cells (Fig. 4B).
(A) Effect of NRG4 isoforms on cell migration of NIH3T3ErbB4CTa cells. Phase contrast images of cells migrating into the wounded area in an in vitro scratch wound healing assay. Exogenous addition of refolded peptide representing the NRG4 isotypes (as indicated) was added to a wounded cell layer of NIH3T3ErbB4CTa cells. High-density phase contrast images represent the wounded monolayer at time 0 and 29 h. Bar represents 300 μM. (B) Time course of cell migration. Cell migration was quantified using relative would density (RWD) (see Material and methods) from data collected at 1-h intervals. The data represent the mean of three wells (with triplicate readings taken from each well, n=9) from a 96-well plate, at each time point. Error bars represent mean±s.d.
Citation: Endocrine-Related Cancer 18, 1; 10.1677/ERC-10-0112
During real-time live imaging, the formation and disappearance of filopodia and lamellipodia at the leading edge of migrating cells could clearly be observed in cells at the wound edge when exposed to NRG4A1/A2 peptide (Fig. 5A and C and Supplementary movies, see section on supplementary data given at the end of this article). In contrast, cells exposed to the NRG4B peptides did not appear to be able to produce functional lamellipodia and did not migrate. Long cytoplasmic processes were observed to form with small flattened cytoplasmic regions at the end (Fig. 5B and D and Supplementary movies, see section on supplementary data given at the end of this article).
Lamellipodia and filopodia formation at leading edge of NRG4A1/A2-stimulated cells. Phase contrast images (taken at 1-h time intervals) of NIH3T3ErbB4CTa cells migrating (arrows indicate direction of movement for each cell) into the wounded area after stimulation with NRG4A1/A2 (A and C) and NRG4B1 (B and D). Lamellipodia (indicated by *) and filopodia (indicated by arrowheads in the inset image) are observed at the leading edge of the cells (C). Cells treated with NRG4B1 show little migration (B) and extend long cytoplasmic processes, which have flattened cytoplasmic regions (indicated by *) (D). Bar represents 20 μM.
Citation: Endocrine-Related Cancer 18, 1; 10.1677/ERC-10-0112
Discussion
We have previously described five alternatively spliced variants of the NRG4 gene, two of which (the A forms) encode transmembrane regions and are transported to the plasma cell membrane. They are not secreted but are retained as transmembrane proteins where they may act in a juxtacrine manner or may be released in soluble forms by a highly regulated proteolytic processing system (Hayes et al. 2008). The B isoforms do not encode a transmembrane sequence and following synthesis are released into the cytoplasm as soluble proteins. In each case, they contain the first two thirds of the EGF domain but differ at their C-termini (Hayes & Gullick 2008). The function of these isoforms in the reducing environment of the cytoplasm, which may not permit them to form the classical disulphide bonding pattern, is currently enigmatic.
We have made polyclonal antibodies that detect all of the NRG4 isoforms (anti-127), the A type isoforms only (anti-123) and antibodies specific for NRG4A1 (anti-128), NRG4A2 (anti-135) or NRG4B3 (anti-134). It was not possible to make antibodies specific for the B1 or B2 isoforms as these only differ by a single amino acid at their C-termini. Here, we have used these reagents to describe the expression of the proteins in normal human adult tissues. It is striking that in certain tissues, e.g. the stomach and the kidney, there are distinct differences in the tissue distribution of the NRG4 isotypes (Supplementary Figure 1, see section on supplementary data given at the end of this article), which is indicative of functional differences.
The NRG4B3 protein is expressed in the endothelial cells of all the tissues investigated (except the brain). It will be interesting to explore this system to see how the protein may function and what its particular role is in endothelial cells. In other cases, antibodies apparently give paradoxical results such as in lymphoid cells where the pan-specific antibody (anti-127) did not stain and the A1-specific antibody (anti-128) did show expression. This may be explained, however, by the release from cells of the extracellular domain containing the 127 epitope and intracellular retention of the other half of the molecule containing the 123 and 128 epitopes. This latter fragment may also be biologically significant as it has been reported that intracellular fragments of NRG1 isoforms are involved in ‘reverse-signalling’ where they are translocated to the cell nucleus and affect gene transcription (Hancock et al. 2008). It is not as yet known if any of the NRG4 isoforms can function in this way, but it is clearly a possibility as NRG4 was detected in the nuclei of salivary glands and spermatocytes (Supplementary text, see section on supplementary data given at the end of this article).
We have reported previously that high levels of NRG4 expression (anti-123) in prostate cancer are associated with advanced tumour stage and described the expression of NRG4A2 mRNA in cell lines derived from metastatic prostate disease and not those derived from localised disease (Hayes et al. 2007). Here, we show that high levels of expression of the NRG4A1 isoform are positively associated with high PSA levels. It is therefore conceivable that NRG4 is influencing disease spread and is associated with advanced tumour stage. However, all the cases of prostate cancer used in this study are designated M=0 (pTMN – pathological TNM), and NRG4A2 (anti-135) is not present in the majority of these cases. Further investigation is therefore required to establish whether NRG4A1 and NRG4A2 are associated with tumour spread.
In the cell migration assays, the refolded NRG4A1/A2, which consists of a full-length EGF domain, stimulated the formation of membrane ruffles, actin reorganisation, and the formation of lamellipodia and filopodia and cell migration. The NRG4B peptides did not induce any of these responses, and surprisingly, less migration was observed than in the untreated control. It is not known whether these forms can bind to ErbB4 but when added to cells they do not appear to stimulate ErbB4 phosphorylation that is induced by NRG4A1/A2 (Hayes et al. 2007). Molecular modelling suggests that, if they were to form a structure similar to a partial EGF domain, it is possible that they could bind to ErbB4 (Hayes & Gullick 2008). One intriguing possibility is that they could act as antagonists rather than agonists, but this hypothesis needs to be tested experimentally.
In summary, the expression of all the known forms of NRG4 is highly correlated such that in some prostate cancers the system is essentially absent, and in others, part or all of the isoforms are produced. We have also shown here that the homologous full-length EGF domain of NRG4A1/A2 in assays of cell architecture and motility induces lamellipodia and filopodia formation and increases motility, but the partial EGF domain found in NRG4B1 and NRG4B3 does not.
Supplementary data
This is linked to the online version of the paper at http://dx.doi.org/10.1677/ERC-10-0112.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
N V L Hayes was supported by the Association of International Cancer Research (AICR), Grant no 20816. E Blackburn is supported by the E.B. Hutchinson Trust, Grant no 20563.
Author contribution statement
Experimental work was carried out by N V L Hayes and E Blackburn. The histological scoring and Gleason grading were performed by M M Boyle and G A Russell. The statistical analyses were done by T M Frost and B J T Morgan. The experimental design was conceived by W J Gullick, and the manuscript was written by N V L Hayes and W J Gullick.
References
Carpenter G & Liao HJ 2009 Trafficking of receptor tyrosine kinases to the nucleus. Experimental Cell Research 315 1556–1566 doi:10.1016/j.yexcr.2008.09.027.
Citri A & Yarden Y 2006 EGF–ERBB signalling: towards the systems level. Nature Reviews. Molecular Cell Biology 7 505–516 doi:10.1038/nrm1962.
Citri A, Skaria KB & Yarden Y 2003 The deaf and the dumb the biology of ErbB-2 and ErbB-3. Experimental Cell Research 284 54–65 doi:10.1016/S0014-4827(02)00101-5.
Cohen BD, Green JM, Foy L & Fell HP 1996 HER4-mediated biological and biochemical properties in NIH3T3 cells. Evidence for HER1–HER4 heterodimers. Journal of Biological Chemistry 271 4813–4818 doi:10.1074/jbc.271.9.4813.
Elenius K, Choi CJ, Paul S, Santiestevan E, Nishi E & Klagsbrun M 1999 Characterisation of a naturally occurring ErbB4 isoform that does not bind or activate phosphatidyl inositol 3-kinase. Oncogene 18 2607–2615 doi:10.1038/sj.onc.1202612.
Falls DL 2003 Neuregulins: functions forms and signalling strategies. Experimental Cell Research 284 14–30 doi:10.1016/S0014-4827(02)00102-7.
Hancock ML, Canetta SE, Role LW & Tammage DA 2008 Presynaptic type III neuregulin1-ErbB signalling targets alpha7 nicotinic acetylcholine receptors to axons. Journal of Cell Biology 181 511–521 doi:10.1083/jcb.200710037.
Harari D, Tzahar E, Romano J, Shelly M, Pierce JH, Andrews GC & Yarden Y 1999 Neuregulin-4: a novel growth factor that acts through the ErbB-4 receptor tyrosine kinase. Oncogene 18 2681–2689 doi:10.1038/sj.onc.1202631.
Hayes NV & Gullick WJ 2008 The neuregulin family of genes and their multiple splice variants in breast cancer. Journal of Mammary Gland Biology and Neoplasia 13 205–214 doi:10.1007/s10911-008-9078-4.
Hayes NV, Blackburn E, Smart LV, Boyle MM, Russell GA, Frost TM, Morgan BJT, Baines AJ & Gullick WJ 2007 Identification and characterisation of novel splice variants of neuregulin 4 in prostate cancer. Clinical Cancer Research 13 3147–3155 doi:10.1158/1078-0432.CCR-06-2237.
Hayes NV, Newsam RJ, Baines AJ & Gullick WJ 2008 Characterisation of the membrane-associated products of the neuregulin 4 gene. Oncogene 27 715–720 doi:10.1038/sj.onc.1210689.
Hynes NE, Horsch K, Olayioye MA & Badache A 2001 The ErbB receptor tyrosine kinase family as signal integrators. Endocrine-Related Cancer 8 151–159 doi:10.1677/erc.0.0080151.
Jones FE 2008 HER4 intracellular domain (4ICD) activity in the developing mammary gland and breast cancer. Journal of Mammary Gland Biology and Neoplasia 13 247–258 doi:10.1007/s10911-008-9076-6.
Jones JT, Akita RW & Sliwkowski MX 1999 Binding specificities and affinities of egf domains for ErbB receptors. FEBS Letters 447 227–231 doi:10.1016/S0014-5793(99)00283-5.
Kainulainen V, Sundvall M, Maatta JA, Santiestevan E, Klagsbrun M & Elenius K 2000 A natural ErbB4 isoform that does not activate phosphoinositide 3-kinase mediates proliferation but not survival and chemotaxis. Journal of Biological Chemistry 275 8641–8649 doi:10.1074/jbc.275.12.8641.
Kita YA, Barff J, Luo Y, Wen D, Brankow D, Hu S, Prigent SA, Gullick WJ & Nicolson M 1994 NDF/Heregulin stimulates the phosphorylation of Her3/erbB3. FEBS Letters 349 139–143 doi:10.1016/0014-5793(94)00644-X.
Kurokawa K & Matsuda M 2005 Localised RhoA activation as a requirement for the induction of membrane ruffling. Molecular Biology of the Cell 16 4294–4303 doi:10.1091/mbc.E04-12-1076.
Montero JC, Roderiguez-Barrueco R, Ocana A, Diaz-Roderiguez E, Esparis-Ogando A & Pandiella A 2008 Neuregulins and cancer. Clinical Cancer Research 14 3237–3241 doi:10.1158/1078-0432.CCR-07-5133.
Rajkumar T, Stamp G, Hughes CM & Gullick WJ 1996 c-erbB-3 protein expression in ovarian cancer. Clinical Molecular Pathology 49 199–202 doi:10.1136/mp.49.4.M199.
Ritch PA, Carroll SL & Sontheimer H 2003 Neuregulin-1 enhances motility and migration of human astrocytic glioma cells. Journal of Biological Chemistry 278 20971–20978 doi:10.1074/jbc.M213074200.
Srinivasan R, Poulsom R, Hurst HC & Gullick WJ 1998 Expression of the c-erbB-4/HER4 protein and mRNA in normal human fetal and adult tissues and in a survey of nine solid tumour types. Journal of Pathology 185 236–245 doi:10.1002/(SICI)1096-9896(199807)185:3<236::AID-PATH118>3.0.CO;2-7.
Stein RA & Staros JV 2006 Insights into the evolution of the ErbB receptor family and their ligands from sequence analysis. BMC Evolutionary Biology 6 79 doi:10.1186/1471-2148-6-79.
Steinthorsdottir V, Stefanson H, Ghosh S, Birgisdottir B, Bjornsdottir S, Fasquel AC, Olafsson O, Stefansson K & Gulcher J 2004 Multiple novel transcription initiation sites for NRG1. Gene 342 97–105 doi:10.1016/j.gene.2004.07.029.
Sundvall M, Iljin K, Kipinin S, Sara H, Kallioniemi OP & Elenius K 2008 Role of ErbB4 in breast cancer. Journal of Mammary Gland Biology and Neoplasia 13 259–268 doi:10.1007/s10911-008-9079-3.
Yarden Y & Sliwkowski MX 2001 Untangling the ErbB signalling network. Nature Reviews. Molecular Cell Biology 2 127–137 doi:10.1038/35052073.
