The synthetic peptide P111-136 derived from the C-terminal domain of heparin affin regulatory peptide inhibits tumour growth of prostate cancer PC-3 cells

Prostate cancer is among the leading malignancies in men throughout much of the industrialized world and ranks second among causes of death from cancer. The lack of not enough effective treatments indicates a need to develop novel treatment strategies targeting new molecules like growth factors. Epithelial-stromal interactions play a pivotal role in the functional integrity of the normal prostate adult gland [1]. This physiological process requires complex interactions between peptide growth factors and growth modulators, which may be regulated either by androgens or by other factors [2, 3]. Any imbalance in these interactions, such as up or down regulation of growth factors or their receptors or a switch from paracrine to autocrine mediation of growth-factor pathways leads to prostate tumour progression. Among the growth-factor families involved in prostate-cancer progression, Transforming Growth Factor-beta (TGFβ), Fibroblast Growth Factors (FGFs), Epidermal Growth Factor (EGF) and heparin affin regulatory peptide (HARP) were reported to play a prominent role [3]. HARP, also called pleiotrophin is a 136 amino acids secreted polypeptide that forms with the protein midkine (MK) a specific family among the heparin-binding growth factors [4]. During embryonic development, HARP is expressed in tissues originating in the mesoderm and neuroectoderm, suggesting a role in epithelium-mesenchyme interactions and in neuronal migration. In adults, HARP expression is limited except at sites such as the mammary gland and uterus associated with reproductive angiogenesis [2]. Furthermore, HARP overexpression has been documented in pathologies associated with cell proliferation and angiogenesis, such as rheumatoid arthritis [5] and tumour growth [6]. HARP has been shown to exert oncogenic potential by transforming various cell lines upon HARP cDNA transfection [7, 8]. In addition, HARP has been shown to play a key role in prostate cancer. Thus, plasma HARP levels were elevated in patients with prostate cancer [9, 10]. Furthermore, HARP protein was associated with epithelial cells in prostate cancer but not in normal prostate tissue and the mRNAs were located in the stromal compartment, suggesting a paracrine mechanism of action for HARP [11]. In vitro, HARP overexpression in normal prostate epithelial PNT-1A cells induced both anchorage-independent and anchorage-dependent growth at low serum concentrations. HARP was also mitogenic for PC-3, LNCaP, and DU145 cell lines [11]. The growth-promoting effect of HARP on prostate cancer cells was also confirmed using an antisense strategy, which established HARP as an important autocrine growth factor for the LNCaP prostate-cancer cell line and as a paracrine factor involved in angiogenesis [12].

Two transmembrane proteins with intracellular catalytic domains have been described as HARP receptors: the receptor-type protein tyrosine phosphatase beta/zeta (RPTP β/ζ) and the anaplastic lymphoma kinase (ALK) receptor. The mitogenic and anti-apoptotic activities of HARP were initially linked to the high-affinity tyrosine kinase receptor ALK in a process mediated by the phosphatidylinositol 3-kinase and MAP kinase signaling pathways [13]. ALK was first identified as a constitutively active, oncogenic, chimeric nucleophosmin-ALK fusion protein [14]. Like HARP, ALK is expressed during normal embryonic development in the similar pattern [15]; however it is also overexpressed in different several human cancers [16, 17]. The neurite outgrowth, the cell migration and adhesion activities of HARP were initially associated with the chondroitin sulfate proteoglycan RPTP β/ζ [18, 19]. HARP was shown to signal through enforced dimerization of RPTPβ/ζ which, in turn, results in a loss of the RPTPβ/ζ catalytic tyrosine phosphatase activity. Inactivation of RPTPβ/ζ by HARP interaction increased tyrosine phosphorylation of each of the substrates of RPTPβ/ζ leading to modification of the cytoskeleton network via the β-catenin, p190RhoGAP, and β-adducin signaling pathway [20‐23]. However, it has also been recently demonstrated that ALK is activated through the HARP/RPTP β/ζ pathway, thus introducing a unique "alternative mechanism of tyrosine kinase receptor activation" [24].

HARP consists of two beta-sheet domains (N- and C-TSR) containing thrombospondin-I repeats, which are flanked by flexible lysine-rich N- and C-terminal tails. Previous studies suggested that the C-TSR domain and the C-terminus regions of HARP were particularly involved in the biological activities of HARP [25‐28]. Thus, the C-terminus, composed of amino acids 111 through 136, was involved in the binding of HARP to the ALK receptor [25, 26]. In these previous studies, mutant HARP protein lacking amino acids 111 through 136 acted as a dominant negative effector of HARP mitogenic, angiogenic, and transforming activities. The synthetic peptide composed of the deleted amino-acid segment P111-136 inhibited HARP mitogenic activity in NIH-3T3, the growth of MDA-MB 231 cells in soft agar and competed with HARP for binding to the extracellular domain of ALK receptor [26] thus representing an interesting pharmacological tool.

The objective of this study was to further complete our investigations on P111-136 and to evaluate its growth inhibitory activity both in vitro and in vivo on the human prostatic adenocarcinoma cell line PC-3.

Previously, it has been shown that peptide P111-136, structurally related to the C-terminal domain of HARP, inhibited specifically the mitogenic activity induced by HARP. It was also shown that P111-136 inhibited the in vitro growth of the human breast cancer cell line MDA-MB 231 suggesting that P111-136 blocked the autocrine stimulation loop of HARP. In order to investigate more precisely the effect of P111-136 on tumour growth, including in vivo study, we investigated the effect of P111-136 on the androgen-independent prostate-cancer cell line PC-3 in which the presence of an autocrine loop of stimulation for this growth factor has been suggested [11, 32].

HARP is expressed in a wide range of tumour cell lines including neuroblastoma, glioblastoma, melanoma and also of cancers derived from the pancreas, breast, and prostate [8, 10, 11, 36‐38]. Numerous studies using ribozyme, RNA interference or antisense strategies [12, 38‐41] showed that HARP was a potential in vivo rate-limiting angiogenic factor in tumour growth and metastasis. As a result, HARP and its two signalling receptors ALK and RPTPβ/ζ are now viewed as promising targets for cancer therapy [17, 38, 39, 42, 43]. Using recombinant deletion mutant of HARP, we previously established that the HARP C-terminus domain, composed of amino acids 111 to 136, was closely involved in the mitogenic, angiogenic and transforming activities of HARP [25, 26]. Thus, the synthetic peptide P111-136 has been shown to bind to the ALK receptor, thereby acting as a dominant inhibitor of the biological activities of HARP. These findings, together with the key role for HARP in prostate-tumour growth, prompted us to further investigate the potential anti-tumour effects of P111-136, using the human androgen-independent adenocarcinoma PC-3 cell line, which expressed HARP and its receptors [11].

In the in vitro experiments reported here, P111-136 inhibited the growth of PC-3 cells. Using whole-cell lysate from PC-3 cells, pull-down experiments with biotinylated P111-136 indicated that P111-136 interact, as expected, with ALK and also with RPTPβ/ζ suggesting that this peptide might bind also to the RPTPβ/ζ receptor. This hypothesis is strengthen by the recent data showing that synthetic peptide including amino acids 122 to 131 derived from the C-terminus region of HARP binds to RPTPβ/ζ [29]. Recently Diamantopoulou et al [33] have demonstrated that this shorter peptide P122-131 was able to inhibit the proliferation of other human prostate cancer cell lines DU145 and LNCaP. It is interesting to mention that these cell lines do not express the ALK receptor (Additional file 1, Figure S1). However, when tested in our conditions, P122-131 inhibited only the anchorage independent growth of PC-3 cells by 22% at 1 μM when compare to 47% for P111-136. The presence of the ALK receptor in PC-3 cells could explain this difference of efficiency between the two peptides but further investigations of both peptides on these different prostate cancer cell lines will be need to clarify this point. When assayed on the human glioblastoma cell line U87 MG described previously as expressing a low level of RPTPβ/ζ [44], P111-136 inhibited the colony formation in a lesser extent (data not shown) supposing that an optimal expression of both ALK and RPTPβ/ζ receptors could be important for a strong effect of P111-136. Therefore, further investigations will be needed to verify this assumption. For several years, the receptors involved in the growth-promoting effect of HARP are a matter of controversy [45] due to the cell system and molecular tools used. However, it has been clearly demonstrated that the binding of HARP or its related protein MK to the ALK receptor activated the intracellular kinase domain and further stimulated the downstream MAP and PI-3 kinase pathways [13, 46, 47]. More, recently Stylianou et al., have also mapped the HARP ligand binding domain on ALK and found that a single chain antibody was able to compete for the HARP binding and inhibit its intercellular downstream signal [48]. Other data have reported that binding of HARP to RPTPβ/ζ receptor oligomerised the receptor and inactivated its intracellular catalytic phosphatase activity, leading to further activation of the Src/Fyn kinase family and β-catenin phosphorylation pathway [20, 21]. The RPTPβ/ζ receptor was found to be involved in HARP-induced cell migration and neurite outgrowth [18, 19] and was then shown to play a role in glioblastoma cell proliferation [42, 43]. In addition, a recent study [24, 42, 43] identified an alternative mechanism in which binding of HARP to RPTPβ/ζ maintained the phosphorylation of ALK, leading to further activation of its pathway. The results described in this study, strongly suggested that a link between ALK and RPTPβ/ζ could occur in PC-3 cells since P111-136 pulled down both receptors.

Nucleolin, a nuclear and cytoplasmic protein initially related to rRNA maturation ribosome assembly [49], was also described as a cell surface protein [35] and as a low affinity binding protein for HARP [34] and MK [50, 51]. Since nucleolin was clearly identified in our pull down experiments, it is tempting to suggest that this protein could also be implicated in the HARP biological activities. As support of this hypothesis, we have shown that anti nucleolin antibody inhibited the biological activity of HARP in MDA-MB 231 cell line [52]. The mechanism of this inhibition remains to be studied. However, previous data have shown that the β-sheet domains of HARP, especially those located on the C-terminal part of the molecule, were involved in the binding of HARP to nucleolin. Since nucleolin was pulled down by P111-136 in our experiments, we could speculate that the peptide did not bind directly to nucleolin but through other non identified molecules, this possibility is currently under evaluation.

The strong inhibition of P111-136 observed on the growth of PC-3 cells in vitro and their tumour xenograft in nude mice could be explain because it targets directly or indirectly different proteins of the HARP pathway. Furthermore in vivo, the inhibition of tumour growth was linked with an increase in PC-3 cells apoptosis and a decreased in the tumour associated angiogenesis. During this treatment, no side effects were observed on mice suggesting that P111-136 did not present any toxicity nor cause additional pathology on animals. Recently Grzelinski et al., [53] have demonstrated an enhanced antitumor effect on glioblastoma using a double ribozyme strategy targeting HARP and one of its receptor. It is tempting to speculate that such results could be obtained on PC-3 tumours in combining P111-136 treatment with another molecule that target the HARP pathway.

Evidence that C-terminus maturation influences the biological effects of HARP was obtained in an earlier study, in which a 14-kDa C-terminal truncated form of HARP influenced the proliferation of cells like bovine epithelial lens cells [54]. Thus, natural processing of the HARP molecule could be pivotal in regulating the biodistribution and biological effects of HARP in health and disease. In keeping with this possibility, we showed that plasmin and MMP-2 cleaved HARP in vitro, releasing various peptides that may differentially affect the angiogenic and mitogenic activities of HARP [55, 56]. GAGs in the microenvironment may protect HARP from this enzymatic cleaving. Thus, the biological effect of HARP is the net result not only of HARP secretion and degradation, but also of specific enzymatic processing, which depends on the proteases and GAGs present in the microenvironment and may generate peptides that have diverse (and perhaps opposite) biological effects.

Our results demonstrate that P111-136 strongly inhibits the proliferative effect of HARP on in vitro and in vivo growth of PC-3 cells. This inhibition of P111-136 could be link to a direct or indirect binding of this peptide to the different HARP interacting proteins (ALK, RPTPβ/ζ, nucleolin). In vivo, the daily P111-136 treatment was as effective as the one of a clinical drug used in prostate cancer therapy. Furthermore the tumour growth inhibition was associated with an inhibition of angiogenesis and an increase in PC-3 cells apoptosis. Thus, P111-136 may be considered as an interesting pharmacological tool to interfere with tumour growth that has now to be evaluated in other cancer types and animals models.