1.  INTRODUCTION

       The IGFBPs are a family of 6 proteins which bind IGF-1 and IGF-2 with high affinity. These proteins can inhibit or enhance IGF action and the same binding protein can have both inhibitory or stimulatory effects dependent upon its concentration, phosphorylation status and proteolysis. Direct effects of some of these IGFBPs (i.e. independent of their ability to bind IGFs) have also been described.In this review we will focus upon the roles of these molecules in controlling cell death during mammary involution at the end of lactation and a possible IGF-independent effect of IGFBP-5 controlling the plasminogen system that is responsible for initiating the remodelling of the extracellular matrix (ECM).

1.1 IGFBPs in development

      Few studies have examined IGFBP expression in the developing rodent mammary gland although both IGFBP-2 and -5 have been implicated in aspects of branching morphogenesis which occurs during pregnancy. Similarities exist however with respect to the developing mammary gland and a variety of embryonic developmental processes and the mammary gland has been described as an ideal tissue to study these processes since its development occurs mainly in the adult. Based upon the observations of Wood et al and our own studies in the involuting mammary gland we have also examined IGFBP-2 and -5 expression in the developing mouse, rat, and chick embryo.Development in the embryonic limb bud shares many features with other systems since epithelial-mesenchymal interactions underlie morphogenesis, and the mammary gland is another example of this. Similarities between the limb bud and the mammary gland are apparent in relation to the IGF system. In the limb bud IGFBP-2 and –5 are expressed in the epithelial cells of the apical ectodermal ridge (AER), while IGF-1 is expressed in the neighbouring subridge mesoderm. This is similar to the mammary gland, where IGFBP-2 and –5 are expressed in mammary epithelial cells while IGF-1 is expressed by the stromal cells. In addition, several other factors shown to be important in limb bud morphogenesis have also been shown to be developmentally regulated in the foetal and postnatal mammary gland, including Msx-1, Msx-2, BMP-2 and BMP-4 .Intriguingly, all of the human IGFBP genes are localized to the same chromosomal regions as different homeobox (Hox) gene families; IGFBP-2 and –5 are linked to the Hox D gene family on human chromosome 2 in opposite  transcriptional  orientations.  The  functional  significance  of thislinkage is unknown, but vertebrate Hox genes are well known for their important functions during embryonic development. Recently, it has been shown that loss of function of Hoxa9, Hoxd9 and Hoxb9 in mice leads to impairment of adult mammary gland development during and after pregnancy. These studies suggest a direct role for these factors in directing differentiation of the mammary epithelium ductal system prior to lactation.

1.2   IGFBP expression in the mammary gland and milk during lactation

        The milk of all species studied contains IGFs and IGFBPs which vary in a species-dependent manner . The presence of IGFs in the milk is similar to that of other growth factors: higher in colostrum than in mature milk.IGFBP-2 is the major IGFBP in human milk with concentrations 10-fold higher in milk than maternal serum suggesting that milk IGFBP-2 is made in the mammary gland ^.Porcine milk contained IGFBP-1, 2, 3 and 4 and some of this IGFBP-3 was in association with ternary complexes IGFBP-2 and -3 mRNA are expressed in porcine mammary gland indicating that they are produced in the mammary gland rather than transported into the gland from blood.Bovine colostrum contains IGFBP-1,-2, -3, and -4 and has high IGF-1 levels although serum IGF-1 is low. IGFBP-3 predominates in prepartum secretions and colostrum but is reduced postpartum.Rat milk contains IGFBP-3, IGFBP-2, and IGFBP-4 and lactating mammary gland expressed IGFBP-2 mRNA suggesting that it is synthesised in the gland.IGFBP-4 expression has been described in rat mammary gland although its concentration in milk is low and we were unable to detect IGFBP-4 expression until involution of the mammary gland. Mammary gland IGFBP-3 mRNA is decreased at parturition and remains low in lactation suggesting that milk IGFBP-3 is serum derived. A portion of the IGFBP-3 in milk was in association with ternary complexes. We were unable to detect IGFBPs-1,-3 or –6 in rat mammary gland during lactation or involution. Although IGFBP-5 and –6 have not been detected in the milk during lactation in any of the species studied, the mammary gland of the bitch expresses IGFBP-2 and IGFBP-5 and progestin administration decreases IGFBP-5 mRNA levels.These studies have demonstrated the species-specific nature of the expression of IGFBPs in milk but, because of their descriptive nature they have failed to shed light on the roles played by the binding proteins at this time. Are they present to transport IGFs, to inhibit the actions of the IGFs or to inhibit the degradation of IGFs in the neonatal gut? It was our intention to ascribe functions to the IGFBPs present in the mammary gland and milk and we therefore examined changes in IGFBP expression induced by involution of the mammary gland as well as examining the hormonal control of IGFBP expression during lactation.

1.3  Changes in IGFBP expression during involution of the mammary gland

       A dramatic increase in the concentration of IGFBP-5 occurs in rat milk within 24 h of mammary involution induced by removal of the suckling young. Increased levels of IGFBP-5 mRNA, as well as more modest increases in IGFBP-2 and -4 (but not IGFBP- 1, -3, and -6) were evident in the mammary gland and in situ hybridisation studies showed that expression occurred in the mammary epithelial cells.IGF-1 has been implicated as a survival factor for many different cell types, including the mammary epithelium and since milk from involuting mammary glands contained large quantities IGFBP-5, we proposed that this served to inhibit IGF-1 actions in the gland. Our studies also showed that both prolactin and GH maintained mammary gland function by inhibiting programmed cell death or apoptosis. We further demonstrated that only prolactin was able to influence IGFBP-5 synthesis serving as a potent inhibitor of IGFBP-5 expression. These findings enabled us to propose a mechanism for the interactive effects of prolactin and GH. GH increases IGF- 1 synthesis whilst prolactin represses IGFBP-5 synthesis thereby maximising the effect of IGF-1 on mammary epithelial cell survival. Several studies have indicated that IGFBP-5 interacts with a variety of components of the ECM. However we demonstrated that IGFBP-5 in milk associates with the casein micelle fraction. The affinity of IGFBP-5 for IGF-1 was approximately 5-8 x 10^-10M, which is higher than

that of the type-I IGF receptor and the concentration of IGFBP-5 in milk was approximately 40 µg/ml. These findings support the proposal that IGFBP-5 acts as an inhibitor of IGF action in milk.

1.4 IGF-independent effects of IGFBPs

      Although all of the IGFBPs bind the IGFs, these molecules also exhibit additional properties, which have led to suggestions that the IGFBPs have biological effects which are independent of the IGFs. Both IGFBP-1 and-2 contain an RGD sequence which is a common feature of molecules which interact with cell-surface integrins. Transfection of cells with IGFBP-1 led to increased rates of cell migration, an effect which was unaffected by the addition of IGF-1 . IGFBP-3 has been shown to inhibit DNA synthesis in human breast cancer cells independently of its ability to bind IGF-1  .Studies in which IGFBP-3 was shown to inhibit growth of a fibroblast cell line derived from a type-1 IGF receptor knock-out mouse were also compelling evidence for such IGF-independent effects.We thus considered whether the high concentrations of IGFBP-5 in milk from involuting mammary glands was present solely to inhibit IGF actions or whether it was also involved in IGF-independent actions of this IGFBP.

1.5 Interactions of IGFBP-5 with the plasminogen system

        IGFBP-5 was first identified in high concentrations in bone where it binds to hydroxyapatite, a crystalline form of calcium phosphate. The casein micelle is composed of casein molecules stabilized by calcium phosphate nanoclusters. The caseins interact through phosphorylated residues, whereas non-phosphorylated proteins such as a-lactalbumin are not incorporated into the micelles. IGFBP-5 has a consensus recognition site for the true mammary casein kinases and it has been reported to be phosphorylated. There is no evidence, however, that IGFBP-5 is bound to calcium phosphate but we have been able to show that IGFBP-5 is involved in a protein-protein interaction with α_s2-casein and, in particular, the dimeric form of this milk protein. Dimeric α_s2-casein has also been shown to bind to plasminogen and tissue-type plasminogen activator (t-PA) which results in enhanced activity of t-PA and the generation of the enzyme plasmin^. Plasmin cleaves a number of pro-enzymes,

such as prostromelysins and procollagenases, and thereby initiates the degradation of the ECM which occurs at the end of lactation. A possible functional interaction of IGFBP-5 and components of the plasminogen system on the casein micelle became a distinct possibility with the observation that IGFBP-5 binds to plasminogen activator inhibitor-I (PAI-1). PAI-1 binds t-PA and inhibits its actions and thus it is conceivable that IGFBP-5 plays a dual role of binding IGF-1 (to inhibit its actions and induce apoptosis) whilst also binding PAI-1 thereby activating t-PA and consequent ECM remodelling (see figure). Such a mechanism would link cell death and ECM degradation, to ensure a highly coordinated tissue-remodelling process. If this were the case we would anticipate that prolactin, by inhibiting IGFBP-5 synthesis, would increase the bioavailability of PAI-1 and suppress plasmin production. We have recently shown that this is indeed the case, since if lactating rats have their litters removed but are treated with prolactin, both t-PA activity and plasmin production are inhibited. GH also inhibited t-PA activity and this could involve IGF-1 since IGF-1 has been shown to increase PAI-1 mRNA expression. We are currently examining these possibilities with in vitro assays for tPA and plasmin activity as well as producing transgenic mice expressing IGFBP-5 on the β-lactoglobulin promoter and developing various IGFBP-5 mutants to examine the effects of exogenous IGFBPs in vitro and in vivo.Does IGFBP-5 co-ordinate activation of cell death and remodelling of the extracellular matrix?

               


2. CONCLUSIONS