Cathepsin D

Proteases are a large and important group of enzymes, with essential roles in a great diversity of biological processes in all living organisms.¹ Proteases are classified according the mechanisms by which they catalyze hydrolysis of peptide bonds within their protein substrates: serine, cysteine, threonine, metallo- and aspartic protease classes are each named for a key functional group instrumental in their chemical mechanisms. A wide variety of proteases have been implicated throughout the development of the mammary gland, but misregulated expression of some of the same proteases has been implicated in the development and progression of breast cancer.^2,3

Understanding the relationship between the physiological functions of particular proteases during development and their pro-tumorigenic functions during cancer is critical for developing appropriately targeted cancer therapeutics. In this issue of Cancer Biology & Therapy, Margaryan et al.⁴ provide important information for bridging this gap for Cathepsin D, a protease expressed in mammary gland development and also misregulated in breast cancer.

Cathepsin D (CatD) is an aspartic protease of the pepsin superfamily. It is constitutively expressed in nearly all cells, where it is trafficked to endosomes and lysosomes, and serves a role as a major acid hydrolase in intracellular protein turnover (reviewed in ref. 5). CatD localization and activation are regulated by multiple posttranslational processing steps (Fig. 1). In breast cancer cells, CatD is often transcriptionally upregulated in a hormonally dependent fashion,⁶ and increased CatD expression has been found to correlate with tumor aggressiveness, metastasis and poor survival.^7,8 CatD has been functionally linked to a number of mechanisms involved in cancer progression, including proliferation, invasion, metastasis and angiogenesis.^7,8 In various breast cancer cell lines, the trafficking of pro-CatD through alternative pathways has been observed, resulting in differing degrees of extracellular secretion vs. lysosomal accumulation (Fig. 1).^9–11

Figure 1.

Processing and activation of CatD. CatD is produced initially as a pre-pro-enzyme of 412 amino acids (aa). A signal peptide of 20-aa is removed during co-translational translocation across the endoplasmic reticulum (ER) membrane, generating proteolytically inactive pro-CatD. In the ER, sugars are attached at two N-linked glycosylation sites, Asn residues 134 and 263 of pre-pro-CatD, corresponding to residues 70 and 199 of mature CatD. In the Golgi complex, mannose residues are phosphorylated, enabling recognition by the cation-independent mannose-6-phosphate receptor (MPR300); this is the primary pathway mediating the sorting of pro-CatD for transport to lysosomes via endosomes in most cell types. In endosomes, unidentified proteases catalyze in trans the proteolytic removal of the 44-aa pro-peptide to generate active single-chain CatD, while in lysosomes, further processing involving cysteine cathepsins B and L generates mature, active double-chain CatD. In some cell types a significant proportion of pro-CatD is secreted extracellularly; it may have noncatalytic extracellular functions and in acidic environments may also undergo autoproteolytic activation via cleavage at an alternative processing site within the pro-peptide, producing active pseudo-CatD.

Although most cell lines in culture secrete mainly the catalytically inactive pro-CatD, CatD recovered from primary breast cancers is largely present in activated forms.¹² Under acidic conditions pro-CatD can auto-activate to generate catalytically active pseudo-CatD,⁵ and it may be that in the acidic milieu of the tumor microenvironment, CatD auto-activation contributes to its tumor-promoting properties. Alternative mechanisms for secreted pro-CatD activation are possible through endocytic reuptake by tumor or stromal cells; while the mechanisms of pro-CatD endocytosis differ between tumor cells and fibroblasts, in both cases the endocytosed pro-enzyme is routed to lysosomes where it undergoes normal activation (Fig. 1).^13,14 A relatively recent wrinkle in elucidation of the tumor-promoting role of pro-CatD is the finding that some of its activities in breast cancer appear to be independent of its catalytic activity: the mitogenic effect of pro-CatD on breast cancer cells is induced equally by wild-type and a catalytically incompetent mutant of pro-CatD, and this mitogenic activity has been localized to a region within the pro-peptide of pro-CatD.^7,8,15,16

An assumption surrounding much of the investigation into the “defect” allowing overproduction and extracellular secretion of pro-CatD or CatD by breast cancer cells is that “normal” cells should not display similar behavior—they should not produce elevated levels of these proteins, nor secrete them. Is this assumption, in fact, correct? Normal human mammary epithelial cells (HMECs), obtained from reduction mammoplasties, secrete minimal levels of proCatD.⁹ However, it should be noted that HMECs reflect an undifferentiated phenotypic state. Given that the physiological purpose of the mammary gland is lactation, understanding of the physiologically normal expression, trafficking and post-translational processing of CatD must encompass the stages of the developing, lactating and involuting mammary gland. In this issue of Cancer Biology & Therapy, Margaryan et al. present a study in which they have used a mouse model of the pregnancy and lactation cycle to define changes in CatD production and processing that coincide with physiological developmental stages.⁴ Such efforts to elucidate mechanisms of CatD regulation and function in normal breast physiology may ultimately generate better understanding of the roots of CatD dysfunction in breast cancer.

The mammary gland undergoes dramatic changes with every cycle of pregnancy and lactation (Fig. 2).^17,18 Early in pregnancy, the ductwork of the mammary epithelium begins to proliferate and expand. Eventually, the epithelium expands to fill the mammary gland, whilst surrounding fat cells dedifferentiate and shrink into small pre-adipocytes.³ In later pregnancy, epithelial cells differentiate further, in preparation for milk production. The branching mammary tree, comprised of ducts and terminal lobules, is lined with a polarized layer of secretory epithelial cells, oriented with their apical surfaces toward a central lumen and their basal surfaces in contact with the supporting basement membrane and myoepithelial cells.^3,18 At parturition (20 days in the mouse), milk secretion is triggered and milk is apically secreted into the lumen. With weaning comes postlactational involution, a multi-stage process of glandular regression and remodeling back to the pre-pregnant state.^3,17,19 In the first, reversible stage, which occurs in days 1–2 after suckling pups are weaned from a lactating mouse, apoptotic epithelial cells are shed into the lumen. In the second, irreversible phase, beginning on day 3, there is dramatic upregulation of many extracellular proteases, accompanied by breakdown of the basement membrane and large-scale glandular collapse. A subsequent third phase of involution has been identified, characterized by remodeling of the mammary stroma and redifferentiation of adipocytes.^3,19 Careful regulation of postlactational involution is critical, as alterations in the progression of this process have been linked to subsequent cancer development.¹⁷

Figure 2.

Mouse mammary gland morphogenesis. Images of whole mounts of fourth inguinal mammary glands illustrate the proliferation and regression of epithelial tissue through the pregnancy/lactation/involution cycle. Mature nulliparous mice have developed a ductal tree that fills the fat pad. Lactating mice show extensive glandular growth and cellular differentiation; this phenotype is rapidly reversed during postlactational involution.

In contrast to the low constitutive levels of CatD production and intracellular, lysosomal retention seen in quiescent, nonlactating mammary epithelial cells, a different picture emerges in the processes of lactation and involution. Pro-CatD is secreted at relatively high levels into human, bovine and rat milk.⁷ Recent studies using primary mammary acini isolated from lactating rats have also identified the selective secretion of activated single-chain CatD from the basal side of the epithelial cells, toward the basement membrane, in a manner influenced by the hormone prolactin.^20,21 Pro-CatD in milk may be activated in the low pH environment of the digestive tract and play a role in infant digestion,⁴ while basally-secreted CatD has been proposed to play a role in the proteolytic processing of prolactin to generate a bioactive fragment with anti-angiogenic and pro-apoptotic properties.^21,22 Active single-chain CatD has been implicated as a mediator of apoptosis in other cell types, as well.^23,24 It appears that the trafficking pathways leading to apical and basal secretion of CatD in mammary epithelial cells are distinct and differentially regulated, but the details of their regulation remain to be elucidated.

The study reported by Margaryan et al. in this issue,⁴ and another recent report by Zaragoza et al.²⁵ use rat and mouse models to extend investigations of CatD in the functional mammary gland through the stages of postlactational involution. The view that begins to emerge is one of highly plastic, highly regulated CatD production, modification, processing and activity. Zaragoza et al.²⁵ have found that CatD mRNA is highly upregulated in the first day of involution, while increases in CatD catalytic activity are somewhat delayed, appearing in days 2–3 of involution. These authors have further identified a novel post-translational modification of CatD appearing during involution, nitration of a specific Tyr residue, that may modestly enhance catalytic activity.²⁵

In this issue, Margaryan et al. have comprehensively tracked mammary CatD protein levels and proteolytic activity in a mouse model from late pregnancy and throughout lactation and involution.⁴ They likewise find CatD protein levels to be strikingly elevated from the initial stages of involution, but for the first three days of involution find primarily a mixture of pro-CatD and active single-chain CatD in the tissues. On day 4 of involution, a large spike in mature, doublechain CatD appears. Measuring levels of CatD activity throughout the lactation/involution cycle, they find a shifting pattern in terms of both constitutive activity (representing intracellularly activated single- and double-chain species) and total activity following in vitro activation by acid treatment. In sum these reports illustrate that with regard to CatD production and activity, the mammary epithelium does not have a single, stable functional state, but is highly plastic. This is relevant to breast cancer: elucidation of the mechanisms by which CatD plasticity is regulated in the normal mammary gland may point the way toward identification of specific regulatory checkpoints that have gone awry under pathological conditions.

Another novel observation of the present report is that the glycosylation pattern of CatD appears to change with lactation and involution status;⁴ the authors suggest that this may hint at an important regulatory axis. The prior literature suggests that changes in glycosylation are unlikely to impact the catalytic activity or substrate specificity of CatD, as a bacterially produced, nonglycosylated form of recombinant CatD has activity and specificity virtually identical to mature human liver CatD toward a variety of peptide substrates.²⁶ However, alterations in glycosylation may very well contribute to altered trafficking and compartmental localization of pro-CatD and CatD, controlling exposure to activating proteases and potential substrates, and thus dictating altered biological functions. Pro-CatD glycosylation has been seen to profoundly affect not only its targeting to lysosomes, but also its extracellular secretion.^27,28 Modulations in CatD glycosylation have been noted in breast cancer cells, and it has been suggested that in this context as well, altered glycosylation may contribute to altered trafficking of CatD.⁹

Important questions remaining about CatD function in the lactation/involution cycle include the extent, regulation and significance of basal secretion. Many proteases are basally secreted in the involuting mammary gland and participate in the breakdown of basement membrane and extracellular matrix (ECM).^17,19,29 They can generate bioactive ECM fragments and activate inflammatory pathways leading to recruitment of stromal cells which collaborate in mammary gland remodeling.^17,19,29 Is CatD secretion part of the program of stromal activation mobilized for mammary gland tissue remodeling during postlactational involution? Such involvement could be highly relevant and highly related to the role of secreted CatD in breast cancer. Proteolytic ECM fragments generated during postlactational involution have been found to possess tumorigenic and metastasis-promoting activities and the elevation of these molecules in the involuting mammary gland has been proposed as one of the factors leading to a transient increase in breast cancer risk following pregnancy, as well as increased metastasis and poorer prognosis in pregnancy-associated breast cancers.^17,29

If, alternatively, the large upsurge in CatD production of involution days 1–3 is retained intracellularly, but routed through an alternative trafficking pathway that delays its final arrival and maturation in the lysosome, it may serve a different, as yet unanticipated role in the early involution process. For a protease such as CatD, location and compartmentalization are critical, not only in a gatekeeping capacity determining exposure to activating proteases, but also in controlling the pool of potential substrates upon which activated CatD might act to exert its biological activities. At the same time, it remains possible that some functional activities of CatD in lactation and involution, as in cancer progression, may be modulated through interactions not dependent upon catalytic activity. Careful delineation of these processes is a critical step, potentially informing future therapeutic strategies targeting the tumor-promoting activities of CatD.

Commentary to: Margaryan Naira V, Kirschmann Dawn A, Lipavsky Alina, Bailey Caleb M, Hendrix Mary JC, Khalkhali-Ellis Zhila. New insights into cathepsin D in mammary tissue development and remodeling. Cancer Biol Ther. 2010 This issue.