Heparin-binding epidermal growth factor-like growth factor functionally antagonizes interstitial cystitis antiproliferative factor via mitogen-activated protein kinase pathway activation


  • Jayoung Kim,

    1. The Urological Diseases Research Center, Children’s Hospital Boston,
    2. Departments of Surgery and Biological Chemistry and Molecular Pharmacology, Harvard Medical School, Boston, MA, and
    Search for more papers by this author
  • Susan K. Keay,

    1. Division of Infectious Diseases, Department of Medicine, the University of Maryland School of Medicine and VA Medical Center, Baltimore, MD, USA
    Search for more papers by this author
  • Michael R. Freeman

    1. The Urological Diseases Research Center, Children’s Hospital Boston,
    2. Departments of Surgery and Biological Chemistry and Molecular Pharmacology, Harvard Medical School, Boston, MA, and
    Search for more papers by this author

Jayoung Kim, Enders Research Laboratories, 1148, 300 Longwood Avenue, Boston, MA 02115, USA. e-mail: ja.kim@childrens.harvard.edu



To delineate the mechanism underlying the potential functional relationship between interstitial cystitis antiproliferative factor (APF) and heparin-binding epidermal growth factor-like growth factor (HB-EGF), as APF has previously been shown to decrease the proliferation rate of normal bladder epithelial cells and the amount of HB-EGF produced by these cells.


APF-responsive T24 transitional carcinoma bladder cells were treated with high-pressure liquid chromatography-purified native APF with or without HB-EGF to determine the involvement of signalling pathways and proliferation by Western blot analysis, p38 mitogen-activated protein kinase (MAPK) and extracellular signal-regulated kinase (Erk)/MAPK assays, and 3-(4,5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide (MTT) assay.


Cyclic stretch induced the secretion of HB-EGF from T24 cells overexpressing the HB-EGF precursor, resulting in enhanced proliferation. T24 cells treated with APF had increased p38MAPK activity and suppressed cell growth, events that were both reversed by treatment with a p38MAPK-selective inhibitor. Activation of Erk/MAPK by HB-EGF was inhibited by APF, and APF did not stimulate p38MAPK in the presence of soluble HB-EGF or when cells overexpressed constitutively secreted HB-EGF. Lastly, APF inhibitory effects on cell growth were attenuated by HB-EGF.


These results indicate that HB-EGF and APF are functionally antagonistic and signal through parallel MAPK signalling pathways in bladder cells.


interstitial cystitis


(soluble) heparin-binding epidermal growth factor-like growth factor


epidermal growth factor (receptor)


IC antiproliferative factor


dimethyl sulphoxide


3-(4,5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide.


Interstitial cystitis (IC) is an idiopathic condition characterized by long-term pelvic symptoms of pain, with increased urinary frequency and urgency [1]. Thinning or ulceration of the bladder epithelial lining in the absence of a UTI is frequently present in IC [2]. Urothelial defects associated with symptoms can be caused by disruptions in urothelial tissue integrity, proliferation, metabolism or function [1]. Although multiple hypotheses have been proposed for the primary cause of IC, the underlying molecular mechanism remains largely undefined. Three biomarkers have been identified in urine from IC patients: the epidermal growth factor receptor (EGFR)/ErbB1/HER1 peptide ligands heparin-binding EGF-like growth factor (HB-EGF) and EGF, and an antiproliferative factor (APF) [3–5]. In patients with IC, APF is present and levels of EGF are significantly increased in urine specimens compared with specimens from unaffected control subjects, while levels of HB-EGF are decreased significantly [5]. APF, an unusual soluble glycosylated peptide related to the membrane receptor frizzled-8, is also synthesized by bladder epithelial cells from patients with IC and has been shown to elicit several responses from this cell type [6–10]. Recently, APF was found to bind to a high affinity cell surface receptor, CKAP4/p63 [11], suggesting that an interaction between APF and this cognate protein may play a role in IC. APF suppresses primary bladder epithelial cell proliferation, increases transcellular permeability by lowering expression of proteins involved in intercellular junctional complexes, and reduces HB-EGF production from primary bladder epithelial cells [8,9,12].

In addition to the activation of EGFR, HB-EGF is also a direct activating ligand for the related tyrosine kinase, ErbB4/HER4. HB-EGF is initially expressed as a transmembrane precursor (proHB-EGF), with the soluble form (sHB-EGF) generated by regulated metalloproteinase-dependent ectodomain shedding. Secreted sHB-EGF activates mitogenic and cell survival functions, while proHB-EGF acts as a juxtacrine factor and also mediates translocation of diphtheria toxin from the cell surface to the cytosol [13–15].

One interesting feature of the existing biomarker data on IC is the potential functional relationship between APF and HB-EGF, IC biomarkers whose levels vary inversely in the urine of patients with IC. APF was shown to decrease HB-EGF production by normal human bladder epithelial cells in vitro, and concentrations of recombinant human HB-EGF similar to those found in normal human urine specimens were shown to abrogate the effects of APF on human bladder epithelial cells [16]. Upon removal of APF from the culture medium, HB-EGF production increased just before recovery of cell proliferation [12]. These results, combined with the lower levels of urine HB-EGF seen clinically in patients with IC, suggest that HB-EGF and APF are functionally antagonistic biopeptides.

The goal of the present study was to delineate the mechanism underlying the potential functional relationship between APF and HB-EGF.


APF was purified from the supernatant of bladder epithelial cells explanted from one patient with IC [16]. Mock APF was prepared using cells from an age-, race- and gender-matched normal control subjected to the same purification procedure. Antibodies against p38MAPK, phospho-p38MAPK, Erk/MAPK and phospho-Erk/MAPK were obtained from Cell Signalling Technology Inc. (Danvers, MA, USA). Anti-HB-EGF antibody was purchased from R & D Systems, Inc. (Minneapolis, MN, USA). SB203580 and dimethyl sulphoxide (DMSO) were purchased from Calbiochem (San Diego, CA, USA).

T24 cells (ATCC, HTB-4) were cultured in McCoy’s 5A containing 10% fetal bovine serum, 1%l-glutamine, and 1% antibiotic/antimycotic solution (all from Sigma) routinely at 37 °C/5% CO2 condition in an humidified incubator. T24 cells were transiently transfected with 2 µg of DNA plasmid vectors (vector only), proHB-EGF (full length HB-EGF construct) or HB-EGFcs (constitutively secretable HB-EGF mutant lacking the transmembrane region) using Nucleofector (Amaxa Inc., Gaithersburg, MD, USA) using the manufacturer’s suggested protocols.

For the in vitro cyclic stretch, T24 cells transiently transfected with a proHB-EGF expression plasmid, and verified to express high levels of the protein, were seeded on type I collagen-coated BioFlex culture plates (Flexcell International, Hillsborough, NC, USA) at a density of 1 × 105 cells. Cells were subjected to cyclic relaxation-stretch the next day. One cycle consisted of 5 s of stretch and 5 s of relaxation (0.1 Hz) and elongation to 20% maximum radial stretch at the periphery of the membrane as previously described [15]. The medium was collected 1 h after cyclic stretch and subjected to pull-down assay using heparin-sepharose beads, followed by Western blot analysis to evaluate HB-EGF secretion levels as described previously [17]. T24 cells transiently transfected with vector plasmid alone were used as a control.

For Western blot analysis, cells were solubilized with whole cell lysis buffer [1% Nonidet P-40; and in mm: 50 Tris pH 7.4, 10 NaCl, 1 NaF, 5 MgCl2, 0.1 EDTA, 1 phenylmethylsulphonyl fluoride; and protease inhibitor cocktail tablet (Roche Diagnostocs Corp., Switzerland)] and centrifuged at 12 000 g for 15 min. The supernatant was assayed for protein concentration and equal amounts of protein were used for Western blotting.

For the p38MAPK assay, T24 cells were preincubated at 37 °C with a specific p38MAPK inhibitor (SB203580) or DMSO for 30 min, followed by treatment with APF for 5 min. Cell lysates were prepared using 1 × lysis buffer provided by the company (p38 MAP Kinase assay kit, Cell Signalling Technology Inc.). After immunoprecipitation with phospho-specific p38MAPK antibody-conjugated beads overnight at 4 °C, samples were incubated in kinase buffer [in mm: 25 Tris at pH 7.5; 5 β-glycerolphosphate; 2 dithiothreitol; 0.1 Na3VO4; and 10 MgCl2] with 200 µm ATP to measure kinase activity using 1 µg recombinant ATF-2 protein as a substrate for p38MAPK. After incubation at 30 °C for 30 min for kinase reaction, samples were boiled in SDS-containing buffer and applied to SDS-PAGE gels to detect the phosphorylation levels of ATF-2 by Western blotting using anti-phospho ATF-2 and anti-ATF-2 antibodies.

For the cell proliferation assay, cells were plated onto 24-well tissue culture plates at a density of 1 × 104 cells/well in standard growth medium. Cells were then serum-starved for 12 h before proliferation analysis. To observe the effect of APF as an antagonist of HB-EGF, 10 ng/mL APF or Mock APF was added to the medium 5 min before treatment with 100 ng/mL HB-EGF. To determine the involvement of p38MAPK activity in APF function, cells were pretreated with 10 µm SB203580 for 1 h, and were incubated with APF-containing medium for the indicated times. Cell viability was determined by uptake of 3-(4,5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide (MTT) as described [17].

For statistical analysis, data were compared using a paired Student’s t-test, with P < 0.05 considered to indicate statistical significance.



Bladder epithelial cells have been shown to respond to cyclic mechanical deformation (stretch) with release of cytokines, growth factors and other soluble molecules, such as ATP [18,19]. We have reported that cyclic stretch induced secretion of sHB-EGF in various cell types, including bladder smooth muscle cells in vitro[20,21]. It has been shown in humans that bladder stretch via hydrodistension changes urinary levels of HB-EGF and APF activity [18]. To begin to understand the possible network relationship between APF and HB-EGF, T24 human bladder cells stably expressing the precursor form of HB-EGF (T24proHB-EGF cells) and vector-only (T24vector cells) were constructed. T24proHB-EGF cells were used in the present experiments because they secrete bioactive, sHB-EGF into the medium in a regulated (inducible) manner, and downstream effects of the growth factor can be monitored [15]. Substantially more sHB-EGF was secreted by the T24proHB-EGF cells under stretch conditions than under non-stretch conditions (Fig. 1), as determined using the pull-down assay of conditioned medium collected after 1 h of cyclic stretch.

Figure 1.

Bladder cell stretch results in secretion of sHB-EGF into the medium. A, T24 bladder cells were transiently transfected with an expression plasmid containing proHB-EGF (T24proHB-EGF) or vector only (T24vec), and were subsequently seeded onto stretch plates. At 48 h after transfection, cells were cyclically stretched (20% elongation, 0.5 Hz) for 1 h. Medium from cells under each set of conditions was collected, and a pull-down assay using heparin-sepharose column was used to measure HB-EGF levels under −/+stretch conditions. Secretion of sHBEGF from T24proHB-EGF or T24vec cells was confirmed by immunoblot using an anti-HB-EGF antibody (CM, conditioned medium).


HB-EGF has been shown to be an autocrine regulator of growth of human bladder epithelial cells [15] and to induce a time- and concentration-dependent phosphorylation of Erk/MAPK, Akt1 and p70S6K, but not p38MAPK in these cells. Western blot analysis showed that HB-EGF treatment of T24 cells transiently stimulated Erk/MAPK phosphorylation, reaching a maximum level after 15 min (Fig. 2A). HB-EGF activated the Erk/MAPK pathway with no discernible effect on Erk/MAPK protein expression (Fig. 2A). Phosphorylation levels of Erk/MAPK were measured in response to HB-EGF in a dose–response format. Western blotting showed marked Erk/MAPK phosphorylation when cells were stimulated with 100 ng/mL HB-EGF for 15 min (Fig. 2B).

Figure 2.

Erk/MAPK activation by recombinant human HB-EGF is inhibited by APF. A, T24 bladder cancer cells were serum starved for 16 h, followed by exposure to 100 ng/mL HB-EGF for the times indicated. Western blot analysis was performed using anti-Erk/MAPK and anti-phospho-Erk/MAPK antibodies. B, After serum starvation, T24 cells were stimulated with various doses of HB-EGF (0, 1, 10, and 100 ng/mL) for 15 min. Cell lysates were prepared and applied to Western blotting. C, T24 cells were treated with 10 ng/mL APF for 5 min, and then incubated with medium containing 100 ng/mL recombinant HB-EGF for another 15 min Western blot analysis was performed to determine Erk/MAPK activation status.

In subsequent experiments, we tested the hypothesis that APF exposure inhibits the largely pro-proliferative signalling cascades induced by HB-EGF. Cells were preincubated in APF-containing medium for 5 min, after which the medium was replaced and cells were stimulated with 100 ng/mL recombinant HB-EGF for 15 min. Remarkably, this brief exposure of cells to APF completely inhibited subsequent stimulation of Erk/MAPK phosphorylation by HB-EGF (Fig. 2C). These findings provide the first direct evidence for inhibitory crosstalk between HB-EGF and APF.


APF has been shown to inhibit cell proliferation of T24 cells via inhibition of cell cycle transit at the G2/M phase and an increase in p53 protein expression [10]. Time-dependence experiments showed that APF treatment activated p38MAPK (Fig. 3A,B). This APF-induced effect is specific within this experimental system because APF did not alter other, alternative signalling pathways (Erk/MAPK, JNK/SAPK or Akt [data not shown]). The level of p38MAPK phosphorylation was elevated within 5 min after APF treatment (Fig. 3A). Kinase activity of p38MAPK was also increased in response to APF as shown in Fig. 3B, and kinase activity was suppressed when cells were preincubated with a specific p38MAPK inhibitor, SB203580, before APF treatment. Cell proliferation rate was reduced by ≈40% after the addition of APF in comparison with the Mock peptide control (Fig. 3B). This growth suppressive effect was completely abrogated by pretreatment of cells with SB203580 showing that p38MAPK is an important mediator of APF-induced cell growth arrest (Fig. 3B).

Figure 3.

Abrogation of APF-induced growth arrest by p38MAPK inhibition. A, T24 cells were stimulated by treatment with HPLC-purified APF. At various time points, whole cell lysates were prepared, and Western blot analysis was performed. Activation of p38MAPK was determined using anti-phospho-p38MAPK and anti-p38MAPK antibodies. B, To determine the involvement of p38MAPK, a specific pharmacological inhibitor of p38MAPK, SB203580, was added 1 h before APF treatment. Medium was changed every 2 days and APF was added to the new medium. Kinase activity of p38MAPK was analysed by measuring ATF-2 phosphorylation levels after immunoprecipitation with phospho-p38MAPK antibody. Cell proliferation was measured at 7 days by MTT analysis (* and **, P < 0.05).


To expand our observations and test directly whether the bladder epithelial cell growth factor HB-EGF antagonizes APF-induced growth arrest, we performed two experiments to assess the effect of exogenous treatment of HB-EGF on p38MAPK activation induced by APF treatment. T24 human bladder cells were stimulated with HB-EGF for 15 min, followed by treatment with APF for the indicated time points. Consistent with the above observations (Fig. 2A), HB-EGF stimulated the Erk/MAPK pathway 15 min after treatment. The purified native APF activated p38MAPK (Fig. 4A) and this activation was completely inhibited by the presence of exogenous HB-EGF (Fig. 4A). Furthermore, as shown in Fig. 4B, APF stimulated p38MAPK activation in vector-only (T24vector) cells; in contrast, APF did not activate p38MAPK in T24 cells constitutively expressing sHB-EGF (T24HB-EGFcs). These data strongly suggest that HB-EGF suppresses APF bioactivity by regulating the p38MAPK signal transduction pathway. To further test our hypothesis that the HB-EGF and APF-induced pathways are antagonistic, T24 cells were pretreated with recombinant HB-EGF for 15 min, followed by incubation in APF-containing medium. When cell proliferation was assessed by MTT cell growth assay, proliferation of control cells treated with APF alone for 7 days was decreased by ≈40% compared with the control. Cells pretreated with HB-EGF and then incubated in APF-containing medium in the absence of HB-EGF showed partial recovery of proliferation in comparison with cells cultured in APF-containing medium alone (Fig. 5).

Figure 4.

HB-EGF antagonizes APF-induced p38MAPK activation. A, T24 cells were serum-starved and stimulated with HB-EGF for 15 min. After removal of medium, cells were treated with 10 ng/mL APF time-dependently. Western blot analysis was used to determine phosphorylation status of p38MAPK at the indicated time points. The protein expression level of p38MAPK was determined by anti-p38MAPK antibody. B, To observe the ability of HB-EGF to suppress APF-induced signalling, T24 cells were transiently transfected with expression constructs containing a constitutively secreted form of HB-EGF (HB-EGFcs) or vector only. After serum starvation for 16 h, cells were treated with APF in the absence of serum. APF increased p38MAPK activation rapidly and transiently. However, T24HB-EGFcs cells did not show this activation after APF treatment.

Figure 5.

Recombinant HB-EGF antagonizes APF-induced growth arrest. T24 cells were serum starved, pretreated with 100 ng/mL HB-EGF for 15 min, after which they were incubated in medium containing 10 ng/mL APF (without HB-EGF) for 7 days. Medium was changed every 2 days and APF replenished. Cell viability was analysed by MTT assay as described in the Materials and methods section (* and **, P < 0.05).


The data presented herein show that HB-EGF is secreted after application of a repetitive mechanical force on bladder-derived epithelial cells, and they confirm that the growth inhibition caused by APF is abrogated by HB-EGF. Cyclic stretch-relaxation resulted in the secretion of HB-EGF into the extracellular space from T24 cells engineered to overexpress the cell surface-localized, precursor form of HBEGF (proHB-EGF), and pretreatment of cells with recombinant HB-EGF before treatment with APF decreased the antiproliferative effect of APF in T24 bladder cells.

The present data also indicate that: (i) in vitro treatment of T24 cells with HB-EGF activates the Erk/MAPK pathway; (ii) APF activates the p38MAPK pathway in these cells, and (iii) both growth inhibition and activation of p38MAPK pathway by APF treatment are suppressed by preincubation of cells with HB-EGF. These data highlight an important biological role for the p38MAPK pathway for mediating APF activity in bladder epithelial cells, and are also consistent with previous papers showing inhibition of Erk/MAPK pathway signalling by p38MAPK pathway activation (or the inverse) [22,23]. In addition, these data suggest the testable hypothesis that the balance between urinary APF and HB-EGF effects on bladder epithelial cells might influence the rate of turnover of cells in the bladder mucosa in patients with IC.

Notably, in some patients with IC, symptoms have improved after bladder distension is performed [24]. Distension, which stretches both bladder smooth muscle and epithelium in vivo, has also been shown to result in increased HB-EGF levels in urine and a reduction of urine APF activity [18]. In previous studies, exogenous treatment of cells with recombinant HB-EGF effectively reversed the effects of APF activity on bladder epithelial cell proliferation [16], consistent with the present results. Differences in the experimental design between the former study (in which cells were simultaneously exposed to APF and HB-EGF for the duration of the experiment) and the present study (in which cells were exposed sequentially to APF or HB-EGF) may explain the slight difference between the two studies in the degree to which HB-EGF could reverse APF-induced inhibition of proliferation.

The role of the MAPK pathway (generally considered to consist of Erk/MAPK, p38MAPK, and JNK/SAPK signalling transduction pathways) in IC is currently undefined, although all three of these signalling systems appear to be important for epithelial, endothelial, and fibroblast cell proliferation. While Erk/MAPK and JNK/SAPK pathways are generally considered to mediate pro-proliferation signals, the p38MAPK pathway is usually associated with inhibition of cell proliferation, and the two pathways often function in opposition to one another to control proliferation [22,23]. For example, during the initial phases of wound healing, corneal epithelial cells migrate but do not proliferate, coincident with p38MAPK activation and Erk/MAPK inhibition [25]. Similarly, suppression of hepatocellular carcinoma cell growth by certain agents is associated with p38MAPK activation and Erk/MAPK inhibition [26]. The Erk/MAPK pathway is activated by HB-EGF through binding to the EGFR tyrosine kinase, and triggers diverse biological responses in bladder cancer cells [17]. Studies using a MAPK-dependent promoter showed that Erk/MAPK activation is also associated with proliferation of normal uroepithelial cells [27]. Enhanced p38MAPK expression levels in the urothelium were also evident in Chernobyl cystitis, a urological disease caused by chronic exposure to low-dose Cs radiation [28], suggesting that alterations in the p38MAPK cascade are early molecular events in the pathogenesis of bladder epithelial cells. In addition, JNK/SAPK expression may also be important for control of uroepithelial cell proliferation, with JNK1 levels decreased in both IC (compared with control) and APF-treated (compared with mock APF-treated) normal bladder epithelial cells [8].

In conclusion, we used an in vitro model to show that mechanical distention induces secretion of sHB-EGF in bladder epithelial cells expressing the precursor protein, proHB-EGF. We also identified a role for the p38MAPK pathway as an important mediator of the antiproliferative effects by APF in bladder epithelial cells, and further showed that HB-EGF can decrease the effects of APF on cell proliferation via stimulation of Erk/MAPK signalling. The present findings are consistent with the hypothesis that bladder distention (or stretching) in vivo is able to increase the urinary levels of endogenous HB-EGF, which may in turn suppress inhibition of bladder epithelial cell proliferation by APF. This hypothesis is also supported by a previous report showing that APF activity may be involved in the regulation of HB-EGF production by bladder epithelial cells [16]. These findings suggest a new mechanism for altering APF-stimulated pathological effects on the bladder wall and provide new insight into possible therapeutic strategies for symptom relief in patients with IC.


We thank Dr Rosalyn Adam for providing advice and for critical reading of the manuscript. This research was supported by NIH grants: R37 DK47556, R01 DK 57691, P50 DK65298 (to M.R.F.) and R01 DK52596 (to S.K.K.), and grants from the Fishbein Family IC Research Foundation/Interstitial Cystitis Association (ICA) and New York Academy of Medicine (NYAM) (to J.K.). J.K. is an American Urological Association Foundation Research Scholar.


None declared.