Targeting a novel bone degradation pathway in ... - Wiley Online Library

Journal of Pathology J Pathol 2016; 238: 120–133 Published online 30 November 2015 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/path.4661

ORIGINAL PAPER

Targeting a novel bone degradation pathway in primary bone cancer by inactivation of the collagen receptor uPARAP/Endo180 Lars H Engelholm,1† Maria C Melander,1† Andreas Hald,1 Morten Persson,2 Daniel H Madsen,3 Henrik J Jürgensen,1 Kristina Johansson,1 Christoffer Nielsen,1 Kirstine S Nørregaard,1 Signe Z Ingvarsen,1 Andreas Kjær,2 Clement S Trovik,4 Ole D Lærum,1,5 Thomas H Bugge,3 Johan Eide6 and Niels Behrendt1* 1

Finsen Laboratory/Biotech Research and Innovation Centre (BRIC), Rigshospitalet and University of Copenhagen, Denmark Department of Clinical Physiology, Nuclear Medicine and PET and Cluster for Molecular Imaging, Rigshospitalet and University of Copenhagen, Denmark 3 Proteases and Tissue Remodelling Section, National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, MD, USA 4 Department of Oncology/Orthopaedics, Haukeland University Hospital, Bergen, Norway 5 Department of Clinical Medicine, Gade Laboratory of Pathology, University of Bergen, Norway 6 Department of Pathology, Haukeland University Hospital, Bergen, Norway 2

*Correspondence to: N Behrendt, Finsen Laboratory, Biotech Research and Innovation Centre (BRIC), Rigshospitalet, University of Copenhagen, Ole Maaløes Vej 5, DK-2200 Copenhagen N, Denmark. E-mail: [email protected] †

These authors contributed equally to this study.

Abstract In osteosarcoma, a primary mesenchymal bone cancer occurring predominantly in younger patients, invasive tumour growth leads to extensive bone destruction. This process is insufficiently understood, cannot be efficiently counteracted and calls for novel means of treatment. The endocytic collagen receptor, uPARAP/Endo180, is expressed on various mesenchymal cell types and is involved in bone matrix turnover during normal bone growth. Human osteosarcoma specimens showed strong expression of this receptor on tumour cells, along with the collagenolytic metalloprotease, MT1–MMP. In advanced tumours with ongoing bone degeneration, sarcoma cells positive for these proteins formed a contiguous layer aligned with the degradation zones. Remarkably, osteoclasts were scarce or absent from these regions and quantitative analysis revealed that this scarcity marked a strong contrast between osteosarcoma and bone metastases of carcinoma origin. This opened the possibility that sarcoma cells might directly mediate bone degeneration. To examine this question, we utilized a syngeneic, osteolytic bone tumour model with transplanted NCTC-2472 sarcoma cells in mice. When analysed in vitro, these cells were capable of degrading the protein component of surface-labelled bone slices in a process dependent on MMP activity and uPARAP/Endo180. Systemic treatment of the sarcoma-inoculated mice with a mouse monoclonal antibody that blocks murine uPARAP/Endo180 led to a strong reduction of bone destruction. Our findings identify sarcoma cell-resident uPARAP/Endo180 as a central player in the bone degeneration of advanced tumours, possibly following an osteoclast-mediated attack on bone in the early tumour stage. This points to uPARAP/Endo180 as a promising therapeutic target in osteosarcoma, with particular prospects for improved neoadjuvant therapy. Copyright © 2015 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.

Keywords: bone cancer; collagen; matrix metalloprotease; matrix degradation; osteosarcoma; uPARAP/Endo180

Received 5 March 2015; Revised 8 September 2015; Accepted 8 October 2015

No conflicts of interest were declared.

Introduction The destruction of bone resulting from bone cancers leads to a pronounced risk of fractures and to severe pain. Therefore, in addition to preventing the actual tumour growth, counteracting tumour-induced bone degeneration is an important goal in its own right [1]. In bone metastases originating from carcinoma at several locations in the body, it is well established that bone destruction results from a ’vicious cycle’ of events in Copyright © 2015 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk

which osteoclasts play a dominant role [2]. These events include a complicated series of signalling reactions in which tumour cells, alone or in conjunction with stromal cells, direct the secretion of osteoclast-stimulating factors, such as RANKL. The osteoclast-mediated bone destruction then leads to the release of TGFβ [3] and other growth factors, which further stimulate tumour growth [4]. In these tumours, the role of the tumour cells in bone matrix destruction is thus indirect, and consequently therapeutic counteraction of bone matrix J Pathol 2016; 238: 120–133 www.thejournalofpathology.com

Bone degradation in primary bone cancer

degeneration is directed against osteoclasts, e.g. by using bisphosphonates or reagents directed against RANKL [5,6]. Less is known about the bone degeneration process associated with sarcomas, such as osteolytic osteosarcoma. Although osteosarcoma is a rare form of cancer, this is a particularly important focus area because the disease most often occurs in the younger population. While it is the metastatic capacity of osteosarcoma that has the most direct impact on survival, the destruction of bone surrounding the primary tumours is a highly critical clinical problem in connection with neoadjuvant therapy. Thus, the primary bone tumours will receive preoperative chemotherapy for many months before the tumour-bearing bone or bone segment is removed [7,8]. Therefore, there is a strong need for additional knowledge, both in this connection and to aid the development of treatment counteracting the morbidity due to inoperable tumours. In accordance with their mesenchymal origin, these primary bone cancers deviate from carcinoma-type metastatic bone cancers in several respects. Although it is likely that early-stage osteosarcomas share the above-mentioned mechanism of osteoclast recruitment to accomplish the initial osteolytic lesions [9], the mechanism of progressive bone breakdown is so far unknown. In this work, we identify a novel bone degradation pathway in primary bone cancer and demonstrate the feasibility of counteracting this process in vivo.

121

nuclei, red fluorescent cells and intralysosomal green fluorescence. The intracellular green fluorescence was quantified using a computerized method, as described in Supplementary materials and methods.

Animal experiments All experiments with mice were performed in accordance with Permission No. 2009/561-1652 from the Danish Animal Experiments Inspectorate. The syngeneic mouse bone tumour model with intrafemoral transplantation of NCTC-2472 sarcoma cells has been described previously [11]. Additional details of this mouse model, as well as procedures for antibody therapy, studies on antibody stability in vivo, microcomputed tomography (microCT) of mouse femurs and studies on the effect of antibody on subcutaneous tumours, are described in Supplementary materials and methods (see supplementary material). In all analyses subsequent to antibody treatment of mice, investigators were unaware of specimen identity with respect to the mouse treatment groups.

Standard procedures The following methods were performed as published, with modifications as described in Supplementary materials and methods (see supplementary material): immunohistochemistry, additional staining methods, microscopy and automated image analysis [12–18], collagen internalization assays [14,19] and western blotting of mouse cells and organs [14,20,21].

Materials and methods Results Human tumour tissue Human tumour tissue specimens obtained from osteosarcoma patients with clinical data listed in Table S1 (see supplementary material) and tissue sections of bone metastases from patients with breast, lung and prostate carcinoma were collected and examined under Permission No. 2011/198-3 from REK Vest, Norway.

Assay for cellular degradation of bone protein in vitro For studies on bone collagen turnover in vitro, bovine bone slices (Code DT-1BON1000-96, Immunodiagnostic Systems, Boldon, UK) were subjected to surface protein fluorescence labelling with the amine-reactive reagent, pHrodo® Green 4-Sulfo-2,3,5,6tetrafluorophenyl (STP) ester (Molecular Probes, ThermoFisher Scientific, USA). Detailed procedures for labelling and for the subsequent steps are supplied in Supplementary materials and methods (see supplementary material). NCTC-2472 cells, transfected to express the red fluorescent protein td-Tomato [10] (see Supplementary materials and methods), were seeded on the labelled bone surface. After 3 days of culture, the cells were subjected to Hoechst staining and examined by confocal microscopy to reveal the cell Copyright © 2015 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk

Low abundance of osteoclasts in advanced-stage osteosarcoma To study the bone degradation process in advanced clinical cases of osteosarcoma, we initially examined the histology of osteolytic lesions in human specimens from this disease, diagnosed from clinical, radiological and histological criteria (see supplementary material, Table S1). Even though these specimens had predominantly osteoblastic characteristics, marked zones of bone destruction were evident in all cases. In these regions, the tumour cells typically formed a contiguous layer in intimate contact with the partially degraded bone matrix (Figure 1A, upper left). These tumour cells were clearly recognizable by standard morphological criteria for malignant cells, i.e. irregular nuclear shape, increased nuclear:cytoplasmic ratio, hyperchromasia and mitoses (Figure 1A, lower left). Although cells with the morphological characteristics of osteoclasts could occasionally be found in the same tumours (Figure 1A, example shown in enlarged field, insert in upper left panel), these cells were scarce in the tumour-dense areas and large bone degradation zones devoid of osteoclasts were always present. This pattern was completely different from that found in secondary bone cancers arising from the dissemination of J Pathol 2016; 238: 120–133 www.thejournalofpathology.com

122

LH Engelholm et al

Figure 1. Histology of bone degradation regions in different types of bone tumours. Designations used in all tissue sections shown: Sarcoma, human osteosarcoma including osteolytic regions; Carcinoma, bone metastasis originating from a breast adenocarcinoma. (A) Sections stained with haematoxylin and eosin (H&E) are shown as low magnification overviews (upper; bars = 250 μm) and with higher magnification of the indicated region (lower; bars = 50 μm); bo, bone; tc, tumour cells; oc, osteoclasts; insert in upper left panel, intratumoural cell with osteoclast morphology in the osteosarcoma specimen (bar = 25 μm); the indicated multinucleated cell is the only cell with this morphology in the current overview. (B) Immunostaining of osteoclasts at the bone–tumour interface, using antibody against CD68 [16]; bars = 250 μm; arrows, CD68-positive cells with osteoclast morphology. (C) Staining for TRAP activity as an osteoclast marker (purple); see supplementary material, Supplementary materials and methods; bars = 250 μm. (D) Counting of CD68-positive cells in the bone degradation regions of several types of bone tumour: in each section, the degradation regions assigned for cell counting were randomly chosen by an investigator unaware of the specimen identity; the following specimens were examined: osteosarcoma (sarcoma; n = 11) and bone metastases originating from carcinomas of breast (n = 7), prostate (n = 6) and lung (n = 9). The number of positive cells is represented relative to the length of bone–tumour interface examined in each section; mean, 25th/75th and 5th/95th percentiles are indicated in the plot; differences indicated were statistically significant; *p = 1.5 × 10−2 ; ***p = 1.9 × 10−4 ; ****p = 5.6 × 10−6 ; Welch ́ s t-test. (E) Immunohistochemical localization of uPARAP/Endo180 and MT1–MMP in the same tumour specimens as shown in (A); bars = 50 μm Copyright © 2015 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk

J Pathol 2016; 238: 120–133 www.thejournalofpathology.com


123

Table 1. Expression of uPARAP/Endo180 and MT1–MMP in tumour cells of human osteosarcoma, determined by imunohistochemistry Patient no 1 2 3 4 5 6 7 8 9 10 11

uPARAP/Endo180

MT1–MMP

++ +++ + + ++ ++ ++ ++ +++ ++ ++

Not performed ++ ++ + ++ ++ +++ + +++ +++ +++

Patient numbering refers to patients with clinical data listed in Table S1 (see supplementary material). +, scattered positive tumour cells seen; ++, many cells positive but also negative tumour cells in between, < 50% positive; +++, a majority or all tumour cells positive.

carcinoma-type cancers elsewhere in the body. In these cases, osteoclast-like cells were abundant in the bone degradation zones and tumour cells were generally well separated from the bone contact area by an intermediate layer of stromal cells (example of a mammary carcinoma bone metastasis shown in Figure 1A, right panels); see further below. To ensure that osteoclasts would not escape detection in the osteosarcoma specimens when using morphological criteria alone, we also stained sections for the occurrence of CD68-positive cells and for TRAP activity as a specific osteoclast marker. This confirmed the very scattered occurrence of osteoclasts in the osteosarcoma specimens (Figure 1B, C, left panels). In contrast, in the mammary carcinoma metastasis, osteoclasts positive for CD68 and TRAP proved abundant at the bone interface (Figure 1B, C, right panels). To evaluate this difference between osteosarcomas and epithelial bone cancers in different tumours and in quantitative terms, we counted osteoclasts at the bone degradation zones of specimens from different bone cancers (Figure 1D). When counting the osteoclasts facing the degraded bone, a very pronounced difference became clear. In the carcinoma-type bone metastases, the bone interface was densely populated with osteoclasts, in accordance with the expected role of these cells in the ’vicious cycle’ of bone destruction. Moreover, the metastases displayed this pattern irrespective of the origin (breast, lung or prostate) of the primary tumour. In contrast, very low osteoclast counts were obtained in the bone degradation regions of the osteosarcoma specimens.

Tumour cell expression of bone-degrading components in osteosarcoma This difference between osteosarcoma and bone tumours of epithelial origin opened the possibility that osteosarcoma cells were directly responsible for an osteolytic process. Interestingly, immunohistochemistry revealed that the osteosarcoma cells were strongly positive for the collagenolytic protease MT1–MMP [22] and Copyright © 2015 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk

Figure 2. Mouse model for sarcoma-mediated bone degradation. (A) Regions of bone degradation in transplanted NCTC-2472 tumours in mice. Upon transplantation into mouse femurs, NCTC-2472 tumours grow invasively, leading to pronounced bone destruction. H&E staining (upper) reveals that the regions of bone degradation (dotted line) are lined with a dense cover of infiltrating sarcoma cells; (centre and lower panels) immunostaining for murine uPARAP/Endo180 and MT1–MMP, respectively, using specific rabbit antibodies; both components are strongly expressed by the tumour cells lining the degraded bone; bars = 20 μm; bo, bone; tc, tumour cells. (B) H&E staining of the same mouse tumour shown in (A), examined at low magnification (upper; bar = 100 μm), with a selected region shown at high magnification (lower; bar = 20 μm): lower panel reveals the occasional occurrence of osteoclasts (arrow); these cells were mostly absent in regions of established tumour growth and extensive bone degradation; bone degradation zone is indicated by a dotted line in the upper panel; bo, bone; tc, tumour cells J Pathol 2016; 238: 120–133 www.thejournalofpathology.com

124

for the endocytic collagen receptor uPARAP/Endo180 [23–25] (Figure 1E, left panels, and Table 1), which participate in a combined extracellular and endocytic/lysosomal pathway of collagen breakdown [19–21,26,27] and cooperate in collagen turnover during normal bone growth [28–31]. We therefore considered these proteins candidate components of an osteosarcoma cell-associated degradation machinery, directed against either intact bone or bone pre-eroded by prior osteoclast activity. The metastatic breast carcinomas showed a strikingly different expression of these components (Figure 1E; see also supplementary material, Table S2). In particular, the carcinoma cells were in all cases negative for uPARAP/Endo180 and, in most cases, also negative for MT1–MMP, although in a few specimens tumour cells displayed focal expression of the protease. In most or all of the specimens, some expression of uPARAP/Endo180 and MT1–MMP was observed in part of the stromal compartment separating the islands of carcinoma cells from the bone matrix, as well as in residual osteoblasts or lining cells in those cases where these cells could be identified. However, in those cases the two components were focally expressed and in most cases only few cells were positive. Consequently, the tumour cell expression of uPARAP/Endo180 and MT1–MMP, as observed in osteosarcoma, was not a general property of all bone cancers.

Mouse model for tumour cell-directed bone degradation recapitulates human osteosarcoma To study the possibility that MMPs and uPARAP/ Endo180 play a role in a sarcoma-mediated bone degradation process, we investigated the properties of an established syngeneic mouse model with transplanted NCTC-2472 cells [32]. After intrafemoral inoculation into C3H/HeN mice, these sarcoma cells grow invasively, with concomitant degradation of the bone surrounding the tumour [11]. It turned out that this model recapitulated the characteristics of the bone degradation zones of the human osteosarcoma specimens observed above. In particular, the mouse tumours showed large, contiguous areas of sarcoma cells in intimate contact with the degraded bone (Figure 2A, dotted line in upper panel). Interestingly, like the human osteosarcoma cells, these cells showed strong expression of uPARAP/Endo180 as well as MT1–MMP (Figure 2A, middle and lower). Furthermore, the degradation zones were largely devoid of osteoclasts (Figure 2A, upper), even though scattered osteoclasts could be identified in other parts of the tumour-bearing bone after an extensive survey (Figure 2B).

Tumour cells used in the mouse osteosarcoma model possess bone-degrading capability in vitro To investigate whether NCTC-2472 cells possess direct bone-degrading capability in the absence of other cell types, we utilized a system with cell culture on bone Copyright © 2015 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk

LH Engelholm et al

slices in vitro. When bone slices were surface-labelled with pHrodo-Green STP ester, a protein-labelling reagent displaying selective fluorescence at a low pH [33], a bone surface was obtained with negligible background fluorescence under the conditions of cell culture (pH 7.4). These bone slices could then be used as a support for the seeding of sarcoma cells. If subject to degradation and endocytosis, the labelled bone protein would be revealed by green fluorescence in the intracellular low-pH compartment of lysosomes, according to the pH-sensitive properties of the fluorophore [33]. To enable the visualization of single cells at a different wavelength, the NCTC-2472 cells to be used in this experiment were transfected to express the red fluorescent protein td-Tomato [10]. After 3 days of culture of NCTC-2472 cells on the labelled bone slices, a strong lysosomal fluorescence signal was indeed obtained (Figure 3A, left, pHrodo channel). Thus, under these conditions, the protein component of the surface of the bone slice was subject to degradation and lysosomal routing. This process was sensitive to inhibition of MMP activity, since addition of the broad-spectrum MMP inhibitor galardin (GM6001) [34] strongly reduced the lysosomal fluorescence. This was evident when studying individual microscopic fields at high magnification (Figure 3B, left) and could also be quantified using low magnification with ∼100 cells/field (Figure 3A, B, right, and Figure 3E). To study the role of uPARAP/Endo180 in the same set-up, we utilized a neutralizing antibody, mAb 5f4, against the murine receptor. This antibody depletes cell surface uPARAP/Endo180, leading to a total block of collagen endocytosis and a gradual depletion of the total cellular pool of the recycling receptor, as shown with several cell types in vitro [14]. In separate experiments (Figure 4) we made sure that this capability of mAb 5f4 was indeed retained with NCTC-2472 cells. Thus, mAb 5f4 could efficiently block the cellular uptake and lysosomal accumulation of added solubilized collagen (Figure 4A, B; see also supplementary material, Figure S2A) and depleted uPARAP/Endo180 protein when added during the culture of these cells (Figure 4C; see also supplementary material, Figure S2B). Importantly, when mAb 5f4 was added to the system with NCTC-2472 cells grown on the pHrodo-Green labelled bone slices, the lysosomal accumulation of fluorescent protein was greatly reduced (Figure 3C). The inhibition obtained with mAb 5f4 was quantitatively equivalent to that obtained with the MMP inhibitor (Figure 3E). Altogether, these studies showed that NCTC-2472 cells do possess a capacity for degradation of bone protein, at least when growing on a pre-eroded surface such as a bone slice. Furthermore, this capacity is dependent on the MMP and uPARAP/Endo180 machinery of the cells.

Targeting uPARAP/Endo180 leads to protection against tumour-mediated bone destruction in vivo The most important questions, however, related to the importance of the proposed degradation mechanism in J Pathol 2016; 238: 120–133 www.thejournalofpathology.com


125

Figure 3. Bone degradation with NCTC-2472 cells in vitro. Bone slices, labelled with pH-sensitive pHrodo-Green fluorophore, were used as the support for culture of td-Tomato-transfected NCTC-2472 cells to enable visualization of cells and bone material at different wavelengths (see Materials and methods). The following reagents were added during cell culture: (A) medium alone; (B) MMP inhibitor galardin (20 μM); (C) neutralizing anti-uPARAP/Endo180 mAb 5f4; (D) irrelevant control mAb (anti-TNP): antibodies were used at 20 μg/ml final concentration. After culture for 3 days, the cells were washed and subjected to Hoechst staining for visualization of nuclei (blue). The cells were examined by confocal microscopy, using high magnification (bars = 10 μm, left panels in A–D) or low magnification (bars = 20 μm, right panels) to enable computerized quantification of fluorescence; green fluorescence represents labelled protein in a low pH (lysosomal) compartment. Each field is shown in the form of the green channel alone (pHrodo channel) as well as the merged image including all three channels (Merge). Z-stacks are shown along the Merge representation of each of the left panels to verify intracellular localization of green fluorescence. (E) Three low-magnification fields were randomly selected from each of the conditions A–D: each field was examined by computerized quantification of lysosomal fluorescence area relative to the number of nuclei present; bar chart shows the values obtained (mean with indication of SD) for each of the four conditions; *differences statistically significant; *p ≤ 0.02; **p ≤ 0.001; NS, not significant; Student’s t-test

Copyright © 2015 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk


126

LH Engelholm et al

Figure 4. mAb 5f4 targets uPARAP/Endo180 and collagen uptake in vitro and in vivo. (A, B) Antibody-mediated blocking of collagen uptake and degradation in NCTC-2472 cells in vitro: the cells were pre-incubated and assayed in the presence or absence of blocking anti-uPARAP/Endo180 mAb 5f4. (A) Cells were presented with Oregon Green-labelled gelatin for endocytosis, followed by confocal microscopy: the confocal micrographs reveal intracellular vesicular accumulation of gelatin (green fluorescence) and the prevention thereof in antibody-treated cells; blue, DAPI staining of nuclei; red, cell surface, labelled with WGA-Alexa 594; bar = 10 μm. (B) Cells were assayed quantitatively for endocytosis of 125 I-labelled collagen type IV: bar chart depicts internalized collagen radioactivity in the presence and absence of antibody. (C) Antibody-mediated depletion of uPARAP/Endo180 in vitro: lysates of NCTC-2472 cells, incubated in the presence or absence of blocking mAb 5f4 (10 μg/ml), were assayed for the expression of uPARAP/Endo180 by western blotting (first two lanes); cell lysates from uPARAP-deficient and wild-type skin fibroblasts were used as negative and positive controls, respectively (last two lanes). (D) Depletion of bone tumour uPARAP/Endo180 in vivo upon systemic treatment of mice with mAb 5f4: mice were inoculated in the right femur with NCTC-2472 cells and treated by intraperitoneal injections with blocking mAb 5f4 or irrelevant control mAb, as detailed in Supplementary materials and methods (see supplementary material). After tumour growth and antibody treatment for 21 days, lungs and tumour-bearing femurs were analysed for uPARAP/Endo180 by western blotting, using radiolabelled anti-uPARAP/Endo180 mAb 2 h9 for band detection; bar chart shows band pixel density after background pixel subtraction

vivo. We therefore decided to target this pathway in the transplanted tumour model in mice. Although targeting might be achieved by blocking MMP activity, problems with functional redundancy and unexpected opposite effects of this type of treatment have been noted in some cancers [35]. However, tumour-associated matrix degradation might also be prevented through interference with the endocytic receptor, since inactivation of the uPARAP/Endo180 gene in the murine PymT mammary tumour model leads to blocking of stroma-mediated collagen turnover [36]. Therefore, we considered uPARAP/Endo180 to be a promising target for prevention of osteosarcoma-mediated bone destruction. We first investigated the in vivo potential of the same neutralizing antibody as used in the cell culture Copyright © 2015 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk

experiments above. Treatment experiments in mice revealed that the antibody was indeed capable of systemic down-regulation of the uPARAP/Endo180 protein level upon intraperitoneal administration (see supplementary material, Figure S2C). Furthermore, it proved well suited for therapy experiments, since it displayed a long half-life in the circulation (see supplementary material, Figure S2D). Most importantly, in the current tumour model the antibody did reach the femoral tumour site, as evidenced by intratumoural uPARAP/Endo180 depletion after administration as above (Figure 4D). Cohorts of C3H/HeN mice were inoculated with NCTC-2472 cells in the right femur and treated with the blocking anti-uPARAP/Endo180 antibody or an irrelevant control (anti-TNP) by intraperitoneal J Pathol 2016; 238: 120–133 www.thejournalofpathology.com


injections twice weekly (for details, see supplementary material, Supplementary materials and methods). After 21 days, the experiment was terminated and each tumour-inoculated femur was examined for bone degradation by microcomputed tomography (microCT) scanning (Figure 5A, left panel), as well as picrosirius red collagen staining (right panel). For quantitative analysis of bone integrity, each tumour-bearing femur was compared with the non-tumour-bearing, contralateral bone. The relative bone mineral density (BMD) was then obtained from the microCT data (Figure 5B). A striking effect of treatment was noted with the uPARAP/Endo180 neutralizing antibody providing pronounced protection against bone destruction (Figure 5C). This appeared to be due to a direct effect on an intrinsic activity of the tumour cells, since microscopic examination of the tumour-bone interfaces in the treated mice (Figure 6A, upper) and in mice receiving the negative control antibody (lower) in both cases still revealed a total dominance of tumour cells (tc). Staining with an independent detection antibody against uPARAP/Endo180 [20] revealed a marked depletion of the receptor in these cells when mice had been treated with the neutralizing antibody (Figure 6B, upper), whereas staining for MT1–MMP revealed pronounced expression in tumours from both groups of mice, with no evident difference (see supplementary material, Figure S3). Furthermore, osteoclasts were still only present in less tumour-dense regions, recapitulating the situation shown above without antibody treatment (Figure 2A, B). To exclude the possibility that the antibody might directly influence tumour growth, without this being related to bone degradation, we set up a separate experiment with NCTC-2472 tumours growing subcutaneously to allow a direct quantification of tumour volume. When mice with these tumours were treated with the uPARAP/Endo180-neutralizing antibody as above, there was no effect on tumour size (Figure 6C). To further exclude that the antibody might influence tumour cell proliferation specifically in the bone compartment, we stained sections from the bone tumour-bearing mice for the Ki67 proliferation marker (see supplementary material, Supplementary materials and methods). Quantitative examination revealed no difference between bone tumours from mice receiving anti-uPARAP/Endo180 or control antibody (Figure 6D). Thus, the effect of the uPARAP/Endo180 antibody was indeed directed specifically against bone destruction.

Discussion This study demonstrates the direct involvement of a tumour cell-resident tissue-degrading machinery mediating bone destruction in osteolytic osteosarcoma and the feasibility of targeting this machinery to counteract the breakdown process by systemic treatment in vivo. Copyright © 2015 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk

127

The novel target component, uPARAP/Endo180, is a central player in a recently established collagen degradation pathway, in which it mediates the endocytic uptake of initial cleavage products resulting from MMP-mediated attack [21]. This uptake event is a rate-limiting step in processes such as collagen clearance in the stromal compartment of malignant breast tumours, where uPARAP/Endo180 deficiency leads to collagen accumulation [36]. Of particular importance, uPARAP/Endo180 is centrally engaged in collagen turnover in the bone matrix, being important for normal bone growth and homeostasis, as shown in mice and cattle [30,37]. Although a complicated expression pattern of uPARAP/Endo180 has been reported for prostate cancers [38,39], this receptor is most often confined to mesenchymal cell types [14,40–42]. However, its role in mesenchymal cancers, such as osteosarcoma, has not been studied previously. An investigation of this question was highly relevant, also in the light of the previously reported, strong expression of uPARAP/Endo180 in cultured osteosarcoma-derived cells [43]. Interestingly, uPARAP/Endo180 turned out to be expressed on tumour cells in all osteosarcoma specimens studied. This expression pattern presented a consistent strong contrast to the samples of bone metastases originating from breast carcinomas, in which we found the tumour cells to be negative in all specimens tested. Even though modified versions of this receptor have been observed [44,45], including protein complexes that might hypothetically escape detection in IHC, our detection system includes stringent demasking procedures to prevent any epitope shielding (see supplementary material, Supplementary materials and methods). Furthermore, our finding of tumour cells being negative in breast carcinoma metastases is consistent with the uPARAP/Endo180 expression pattern in primary breast carcinomas, determined by IHC as well as in situ hybridization [42]. The carcinoma cells were also mostly negative for the membrane-anchored collagenolytic MMP, MT1–MMP, although a few positive cases were found (see supplementary material, Table S2). This is consistent with the variable expression of MT1–MMP in primary breast carcinomas observed in other studies [46]. The pronounced difference between the expression patterns of osteosarcoma and breast carcinoma-derived bone metastases indicates that the targeting of tumour-expressed uPARAP/Endo180, as performed in this study, cannot necessarily be generalized to other types of bone tumour. Nonetheless, some expression of uPARAP/Endo180 as well as MT1–MMP was observed in the stromal compartment of the carcinoma metastases, separating the tumour islands from the bone surface (see supplementary material, Table S2). Consequently, targeting studies in additional bone tumour models, including models of bone metastasis, might still be highly relevant, irrespectively of the cellular origin (tumour or stroma) of the receptor. J Pathol 2016; 238: 120–133 www.thejournalofpathology.com

128

LH Engelholm et al

Figure 5. Functional blocking of uPARAP/Endo180 inhibits bone degradation in osteosarcoma. (A) Representative examples of tumour-bearing bone after systemic treatment of tumour-inoculated mice with anti-uPARAP/Endo180 mAb 5f4 (blocking mAb) or irrelevant control mAb: (left) 3D reconstructed surface representations derived from microCT scans of affected bones after treatment; white arrows, position of tumour cell inoculation; each picture includes the contralateral (non-tumour bearing) bone for comparison; (right) sections of the same tumour-bearing bones, stained with picrosirius red for visualization of bone collagen; bar = 2 mm. (B) Procedure for quantification of bone integrity: (upper panel) a 3D volume-rendering model was made from the microCT scanning data, using OsiriX; a working region, common for all bones measured, was selected, based on a fixed distance (90 mm) from the femoral neck; the excluded proximal femur region is shown in grey-scale; (centre panel) regions of interest (ROIs) in tumour bearing bone (Cancer) and in the contralateral femur (Contralateral) were first defined on a single cross-section (pixel threshold value set to exclusively include bone structures); (lower panel) threshold areas were then expanded to include all coherent bone in the working region of each femur. ROI volumes were visualized to ensure that all relevant bone structures were included in the later BMD calculations; the BMD was calculated as the cumulative density values of all pixels in the volume defined by the ROI selection; for each mouse, the relative bone integrity of the tumour-bearing site was calculated as the ratio between the BMD of the tumour-bearing ROI and the BMD of the ROI of the contralateral femur. (C) Quantification of the effect of treatment: the relative BMD values are shown for the two groups of mice, treated with blocking anti-uPARAP/Endo180 mAb 5f4 or irrelevant control mAb (n = 11 in each group): mean, 25th/75th and 5th/95th percentiles are indicated in the plot. Treatment with the anti-uPARAP/Endo180 antibody led to a marked protection of the tumour-bearing bones, compared with mice treated with the control antibody (less reduction in the BMD; p = 0.002, Welch’s t-test)




129

H&E

Tumour

Tumour

Figure 6. Antibody treatment is directed against sarcoma cell-mediated bone degradation. (A) H&E-stained sections of the tumour bone degradation region in mice treated with blocking anti-uPARAP/Endo180 mAb 5f4 (upper) or control mAb (lower). The interface consists of a dense layer of tumour cells (tc), with osteoclasts being absent from this region; bo, bone; bars = 20 μm. (B) Bone tumour sections immunostained for uPARAP/Endo180 with an independent polyclonal detection antibody: little or no expression of uPARAP/Endo180 could be detected in tumours treated with blocking anti-uPARAP/Endo180 mAb (upper), whereas uPARAP/Endo180-positive tumour cells are abundant in mice treated with control antibody (lower; closed arrows); bars = 20 μm; bo, bone; tc, tumour cells. (C) Subcutaneous inoculation of NCTC-2472 cells and determination of tumour growth in mAb-treated mice: tumour volume was measured at 2 day intervals and plotted for each treatment group over time (left) and at the termination of the experiment after 20 days (right); no difference could be detected between mice treated with control mAb (filled circles, n = 10) or blocking anti-uPARAP/Endo180 mAb (squares, n = 10). (D) Quantification of tumour cell proliferation in mAb-treated bone tumours: sections from the tumour-bearing bones of antibody-treated mice were immunostained for cell proliferation marker Ki67 (upper panels); bars = 50 μm. The overlay (centre panels) shows computerized identification of the Ki67-positive area (blue); bar charts (lower panel) represent the computerized data obtained from whole scans of the tumour-bearing bone regions of mice treated with anti-uPARAP/Endo180 mAb 5f4 (Blocking mAb) and control mAb, respectively (n = 6 for each group); no significant difference was noted in the Ki67-positive area (left chart) or number of Ki67-positive cells/mm2 (right chart) in sections from the two groups of mice (Student’s t-test); charts represent mean values with indications of SDs. (E) Models depicting different pathways of bone collagen degradation: (left panels) in epithelial bone metastases, a series of mutually stimulating cellular signalling events leads to osteoclast (oc) activation and cathepsin K-mediated collagen degradation; this ’vicious cycle’ mechanism involves tumour cells and osteoblasts in addition to the bone-degrading osteoclasts [2] (upper left); as a consequence, bone degradation occurs without direct contact between tumour cells and bone (lower left). An equivalent mechanism has been suggested to be involved in early stages of osteolytic osteosarcoma [9]; (right panels) tumour cell-resident collagen degradation mechanism of primary bone cancer, depending on MMP activity and uPARAP/Endo180-mediated endocytosis; our data indicate that this is an important mechanism of bone destruction in advanced stages of osteolytic osteosarcoma, which can be targeted by inactivation of uPARAP/Endo180; in this mechanism, tumour cells are in intimate contact with the partially degraded bone (lower right) Copyright © 2015 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk


130

The pronounced tumour cell expression of uPARAP/Endo180 as well as MT1–MMP in osteosarcoma and the strong protective effect of the neutralizing anti-uPARAP/Endo180 antibody in our mouse model point to a critical role of the receptor in bone destruction in this disease. Studies with osteosarcoma cells and other cell types in vitro suggest an involvement of this receptor in several processes, including matrix turnover as well as cell adhesion and cell migration [20,26,43]. However, the effect of the antibody in our mouse model appears to be directed exclusively against bone matrix degradation (Figures 5, 6). This points to the well-characterized function of uPARAP/Endo180 in collagen internalization and degradation as the responsible molecular mechanism [14,19,20,47]. The details of the balance between the bone degradation mechanism examined here and the osteoclast-dependent ’vicious cycle’ mechanism are not known; see Figure 6E. Although the regions of bone destruction were largely devoid of osteoclasts in our osteosarcoma specimens, our results do not clarify whether sarcoma-mediated bone degradation is actually independent of osteoclast activity during the whole course of the degradation process. The importance of osteoclasts in osteosarcoma-directed bone degradation has been discussed in early investigations, including studies using the same tumour model as employed here [48]. In those studies, it was correctly pointed out that a lack of osteoclasts in a certain lesion does not exclude that these cells had been present at an earlier point [48]. However, for advanced osteosarcomas, it seems unlikely that all 11 of our specimens would display a uniform cellular pattern (Figure 1D) without this reflecting a general scarcity of osteoclasts in the bone degradation zones at that stage of the disease. These specimens were taken with no preference for the current field of research and were just examples of advanced-stage osteosarcoma lesions with ongoing bone destruction. Furthermore, a pronounced down-regulation of osteoclasts in osteosarcoma progression has also been noted by others [49], even though extensive bone degradation is also ongoing in advanced stages of this disease. All in all, it seems likely that the degradation processes mediated by osteoclasts and sarcoma cells both operate in osteosarcoma, depending on the stage of disease and on individual variation. Our studies in vitro with a pre-eroded bone surface in the form of surface-labelled bone slices showed that sarcoma cells can indeed degrade the protein component of this matrix. A possible scenario in vivo would include an initial osteoclast-mediated attack on the bone surface, followed by a sarcoma-mediated, extended degradation of the initially eroded matrix. This would fit well with the tumour cell-resident mechanism being dominant in established tumours with ongoing bone destruction. Indeed, it has been hypothesized previously that osteoclasts play their most important role during early osteosarcoma growth [9]. Furthermore, this combined model could explain the reported observations that a complete lack of osteoclasts in the host mouse leads to failure of bone destruction in Copyright © 2015 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk

LH Engelholm et al

the same osteosarcoma model as used in our study [48], whereas the same tumours can efficiently degrade the bone of a host mouse with a strongly reduced, but still detectable, osteoclast level [50]. As to the composite processes that lead to bone degradation, it is widely accepted that osteoclast activity is essential for efficient dissolution of the inorganic component of the bone matrix in the healthy organism, due to the formation of a low pH compartment in the resorption pit formed by these cells [51]. Whether this is also the sole mechanism of bone demineralization in osteosarcoma is not known. One possibility is that an initial process of bone erosion by osteoclasts is sufficient to create denuded collagen, which is then subject to sarcoma-mediated degradation, as outlined above. Alternatively, a tumour-associated degradation process might proceed during certain disease stages where low pH conditions in the tumour microenvironment [52] could contribute to the dissolution of the inorganic bone component. At a microscopic scale, it is not unlikely that a similar interplay between the current pathway and osteoclast activity is involved in the microresorption processes occurring during normal bone remodelling. Thus, osteoclast deficiency is known to lead to severe skeletal defects [53] but such defects also occur as the result of deficiency of MT1–MMP and/or uPARAP/Endo180 in the osteoblastic compartment [28–30]. In both cases this is ascribed to failure of bone resorption during bone growth. Electron microscopy studies have convincingly demonstrated the involvement of both osteoclasts and osteoblast-related lining cells in bone resorption. Indeed, the latter cells serve to enwrap and digest bone collagen in MMP-dependent processes, both in the case of cleaning resorption lacunae demineralized by osteoclasts in normal bone remodelling and under conditions such as pycnodysostosis, in which the osteoclasts lack cathepsin K [54]. Interestingly, in the latter study [54] it is suggested that a digestion step, removing non-mineralized collagen protruding from the bone surface, precedes the osteoclast attachment step in bone resorption; this step is mediated by osteoblast-like bone lining cells and is dependent on MMP activity. The activity of osteosarcoma cells found in our study is in line with these observations in non-malignant conditions. However, in addition the latter studies open the possibility that non-malignant osteoblasts/lining cells could also contribute to the complicated scenario of bone degradation in osteosarcoma, and possibly even in other bone cancers. It is noteworthy that lining cells are strongly positive for uPARAP/Endo180 in healthy bone [55] and that residual non-malignant osteoblasts/lining cells with focal expression of uPARAP/Endo180 could be identified in four of seven breast carcinoma metastases (see supplementary material, Table S2) and nine of 11 osteosarcoma specimens in our study (see supplementary material, Table S3). Consequently, osteoblast-like cells may indeed be included in the cell population that is functionally targeted in the current therapeutic strategy. The collagen metabolism J Pathol 2016; 238: 120–133 www.thejournalofpathology.com


of non-malignant osteoblasts has been reported to be influenced by uPARAP/Endo180 expression, and the stromal/osteoblastic expression level has been found to be low in the area of larger tumour foci in bone metastases from prostate cancer [39]. In our study, however, a comparison with the osteoblasts and lining cells of normal bone was not possible, since our clinical material comprised exclusively tumour-affected bone tissue. This was the case for both the osteosarcomas and the breast carcinoma metastases, thus precluding an evaluation of any difference of the role of these cells in normal bone, osteosarcoma and carcinoma bone metastases, respectively. Altogether, our findings point to the novel pathway being an important component in a composite system of bone tissue destruction in primary bone cancer. We believe that the targeting strategy described here, complementing those directed against previously established pathways, can be utilized to create novel therapeutic options. In particular, it is possible that the targeting of uPARAP/Endo180 could be conveniently combined with bisphosphonate treatment for the inactivation of osteoclasts, which might also add the advantages of direct antitumour and anti-angiogenic effects of the latter reagents [56,57]. An improved treatment strategy along this line would be relevant, not least in connection with neoadjuvant therapy [7]. A reduction in fracture risk during the preoperative period of chemotherapy is crucially important, and reduced bone destruction might even enhance the effect of simultaneously administered antitumour drugs.

Acknowledgements The excellent technical assistance of Katharina H Stegmann, Lotte B Frederiksen, Lene K Callesen and Bendik Nordanger is gratefully acknowledged. Dr Sanjiv S Gambhir is thanked for the generous gift of the pFU-L2T vector. Senior Statistician Ib Jarle Christensen is thanked for data evaluation and statistical advice. This study was supported by the Danish Cancer Society, the Danish Medical Research Council, the Danish National Advanced Technology Foundation, the Danish Cancer Research Foundation, the Lundbeck Foundation, the Novo Nordisk Foundation, the AP Moller Foundation, the Danish National Research Foundation (Danish–Chinese Centre for Proteases and Cancer), the Intramural Research Programme of the National Institutes of Health (NIH), NIDCR, USA, the European Community’s Seventh Framework Programme FP7/2007-2011 (Grant No. 201279) and the Grosserer Alfred Nielsen og Hustrus Foundation. DHM and HJJ received personal fellowship grants from Copenhagen University Hospital.

Author contributions LHE, MCM and NB planned the study, developed the hypotheses and designed the investigation; LHE, MCM, Copyright © 2015 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk

131

AH, MP and JE performed the experimental work and pathological investigations, with SZI, KSN, HJJ, KJ, CN and NB taking part in specific experiments; DHM, AK, CST, ODL and JE contributed essential reagents, technology and pathological specimens for the study; LHE, MCM, JE and NB interpreted the results; DHM and THB contributed special expertise in bone biology, and CST, ODL and JE special expertise in bone cancer; LHE, MCM and NB wrote the paper, with contributions from all authors.

References 1. Mundy GR. Metastasis to bone: causes, consequences and therapeutic opportunities. Nat Rev Cancer 2002; 2: 584–593. 2. Mundy GR. Mechanisms of bone metastasis. Cancer 1997; 80: 1546–1556. 3. Korpal M, Yan J, Lu X, et al. Imaging transforming growth factor-β signaling dynamics and therapeutic response in breast cancer bone metastasis. Nat Med 2009; 15: 960–966. 4. Sturge J, Caley MP, Waxman J. Bone metastasis in prostate cancer: emerging therapeutic strategies. Nat Rev Clin Oncol 2011; 8: 357–368. 5. Wilson C, Holen I, Coleman RE. Seed, soil and secreted hormones: potential interactions of breast cancer cells with their endocrine/ paracrine microenvironment and implications for treatment with bisphosphonates. Cancer Treat Rev 2012; 38: 877–889. 6. Brown JE, Cook RJ, Lipton A, et al. Serum lactate dehydrogenase is prognostic for survival in patients with bone metastases from breast cancer: a retrospective analysis in bisphosphonate-treated patients. Clin Cancer Res 2012; 18: 6348–6355. 7. Ferrari S, Smeland S, Mercuri M, et al. Neoadjuvant chemotherapy with high-dose Ifosfamide, high-dose methotrexate, cisplatin, and doxorubicin for patients with localized osteosarcoma of the extremity: a joint study by the Italian and Scandinavian Sarcoma Groups. J Clin Oncol 2005; 23: 8845–8852. 8. Gill J, Ahluwalia MK, Geller D, et al. New targets and approaches in osteosarcoma. Pharmacol Ther 2013; 137: 89–99. 9. Endo-Munoz L, Evdokiou A, Saunders NA. The role of osteoclasts and tumour-associated macrophages in osteosarcoma metastasis. Biochim Biophys Acta 2012; 1826: 434–442. 10. Liu H, Patel MR, Prescher JA, et al. Cancer stem cells from human breast tumors are involved in spontaneous metastases in orthotopic mouse models. Proc Natl Acad Sci USA 2010; 107: 18115–18120. 11. Hald A, Hansen RR, Thomsen MW, et al. Cancer-induced bone loss and associated pain-related behavior is reduced by risedronate but not its phosphonocarboxylate analog NE-10790. Int J Cancer 2009; 125: 1177–1185. 12. Sulek J, Wagenaar-Miller RA, Shireman J, et al. Increased expression of the collagen internalization receptor uPARAP/Endo180 in the stroma of head and neck cancer. J Histochem Cytochem 2007; 55: 347–353. 13. Madsen DH, Jurgensen HJ, Ingvarsen S, et al. Endocytic collagen degradation: a novel mechanism involved in protection against liver fibrosis. J Pathol 2012; 227: 94–105. 14. Madsen DH, Ingvarsen S, Jurgensen HJ, et al. The non-phagocytic route of collagen uptake: a distinct degradation pathway. J Biol Chem 2011; 286: 26996–27010. 15. Junqueira LC, Bignolas G, Brentani RR. Picrosirius staining plus polarization microscopy, a specific method for collagen detection in tissue sections. Histochem J 1979; 11: 447–455. 16. Athanasou NA, Puddle B, Quinn J, et al. Use of monoclonal antibodies to recognise osteoclasts in routinely processed bone biopsy specimens. J Clin Pathol 1991; 44: 664–666. J Pathol 2016; 238: 120–133 www.thejournalofpathology.com

132

17. Atkinson SM, Usher PA, Kvist PH, et al. Establishment and characterization of a sustained delayed-type hypersensitivity model with arthritic manifestations in C57BL/6 J mice. Arthritis Res Ther 2012; 14: R134. 18. Scholzen T, Gerdes J. The Ki-67 protein: from the known and the unknown. J Cell Physiol 2000; 182: 311–322. 19. Kjoller L, Engelholm LH, Hoyer-Hansen M, et al. uPARAP/endo180 directs lysosomal delivery and degradation of collagen IV. Exp Cell Res 2004; 293: 106–116. 20. Engelholm LH, List K, Netzel-Arnett S, et al. uPARAP/Endo180 is essential for cellular uptake of collagen and promotes fibroblast collagen adhesion. J Cell Biol 2003; 160: 1009–1015. 21. Madsen DH, Engelholm LH, Ingvarsen S, et al. Extracellular collagenases and the endocytic receptor, urokinase plasminogen activator receptor-associated protein/Endo180, cooperate in fibroblast-mediated collagen degradation. J Biol Chem 2007; 282: 27037–27045. 22. Hotary KB, Allen ED, Brooks PC, et al. Membrane type I matrix metalloproteinase usurps tumor growth control imposed by the three-dimensional extracellular matrix. Cell 2003; 114: 33–45. 23. Behrendt N, Jensen ON, Engelholm LH, et al. A urokinase receptor-associated protein with specific collagen binding properties. J Biol Chem 2000; 275: 1993–2002. 24. Sheikh H, Yarwood H, Ashworth A, et al. Endo180, an endocytic recycling glycoprotein related to the macrophage mannose receptor is expressed on fibroblasts, endothelial cells and macrophages and functions as a lectin receptor. J Cell Sci 2000; 113: 1021–1032. 25. Melander MC, Jurgensen HJ, Madsen DH, et al. The collagen receptor uPARAP/Endo180 in tissue degradation and cancer [review]. Int J Oncol 2015. 26. East L, McCarthy A, Wienke D, et al. A targeted deletion in the endocytic receptor gene Endo180 results in a defect in collagen uptake. EMBO Rep 2003; 4: 710–716. 27. Jurgensen HJ, Johansson K, Madsen DH, et al. Complex determinants in specific members of the mannose receptor family govern collagen endocytosis. J Biol Chem 2014; 289: 7935–7947. 28. Holmbeck K, Bianco P, Caterina J, et al. MT1–MMP-deficient mice develop dwarfism, osteopenia, arthritis, and connective tissue disease due to inadequate collagen turnover. Cell 1999; 99: 81–92. 29. Wagenaar-Miller RA, Engelholm LH, Gavard J, et al. Complementary roles of intracellular and pericellular collagen degradation pathways in vivo. Mol Cell Biol 2007; 27: 6309–6322. 30. Fasquelle C, Sartelet A, Li W, et al. Balancing selection of a frame-shift mutation in the MRC2 gene accounts for the outbreak of the crooked tail syndrome in Belgian blue cattle. PLoS Genet 2009; 5: e1000666. 31. Engelholm LH, Nielsen BS, Netzel-Arnett S, et al. The urokinase plasminogen activator receptor-associated protein/endo180 is coexpressed with its interaction partners urokinase plasminogen activator receptor and matrix metalloprotease-13 during osteogenesis. Lab Invest 2001; 81: 1403–1414. 32. Gu X, Zheng Y, Ren B, et al. Intraperitoneal injection of thalidomide attenuates bone cancer pain and decreases spinal tumor necrosis factor-α expression in a mouse model. Mol Pain 2010; 6: 64. 33. Flores JA, Balseiro-Gomez S, Cabeza JM, et al. A new role for myosin II in vesicle fission. PLoS One 2014; 9: e100757. 34. Almholt K, Juncker-Jensen A, Laerum OD, et al. Metastasis is strongly reduced by the matrix metalloproteinase inhibitor Galardin in the MMTV–PymT transgenic breast cancer model. Mol Cancer Ther 2008; 7: 2758–2767. 35. Coussens LM, Fingleton B, Matrisian LM. Matrix metalloproteinase inhibitors and cancer: trials and tribulations. Science 2002; 295: 2387–2392. 36. Curino AC, Engelholm LH, Yamada SS, et al. Intracellular collagen degradation mediated by uPARAP/Endo180 is a major pathway of Copyright © 2015 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk

LH Engelholm et al

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48. 49.

50. 51. 52. 53.

54.

55.

56.

extracellular matrix turnover during malignancy. J Cell Biol 2005; 169: 977–985. Madsen DH, Jurgensen HJ, Ingvarsen S, et al. Differential actions of the endocytic collagen receptor uPARAP/Endo180 and the collagenase MMP-2 in bone homeostasis. PLoS One 2013; 8: e71261. Kogianni G, Walker MM, Waxman J, et al. Endo180 expression with cofunctional partners MT1–MMP and uPAR–uPA is correlated with prostate cancer progression. Eur J Cancer 2009; 45: 685–693. Caley MP, Kogianni G, Adamarek A, et al. TGFβ1–Endo180dependent collagen deposition is dysregulated at the tumour–stromal interface in bone metastasis. J Pathol 2012; 226: 775–783. Behrendt N. The urokinase receptor (uPAR) and the uPAR-associated protein (uPARAP/Endo180): membrane proteins engaged in matrix turnover during tissue remodeling. Biol Chem 2004; 385: 103–136. Engelholm LH, Ingvarsen S, Jurgensen HJ, et al. The collagen receptor uPARAP/Endo180. Front Biosci (Landmark Ed) 2009; 14: 2103–2114. Schnack Nielsen B, Rank F, Engelholm LH, et al. Urokinase receptor-associated protein (uPARAP) is expressed in connection with malignant as well as benign lesions of the human breast and occurs in specific populations of stromal cells. Int J Cancer 2002; 98: 656–664. Sturge J, Wienke D, Isacke CM. Endosomes generate localized Rho–ROCK–MLC2-based contractile signals via Endo180 to promote adhesion disassembly. J Cell Biol 2006; 175: 337–347. Rodriguez-Teja M, Gronau JH, Breit C, et al. AGE-modified basement membrane cooperates with Endo180 to promote epithelial cell invasiveness and decrease prostate cancer survival. J Pathol 2015; 235: 581–592. Rodriguez-Teja M, Gronau JH, Minamidate A, et al. Survival outcome and EMT suppression mediated by a lectin domain interaction of Endo180 and CD147. Mol Cancer Res 2015; 13: 538–547. Tetu B, Brisson J, Wang CS, et al. The influence of MMP-14, TIMP-2 and MMP-2 expression on breast cancer prognosis. Breast Cancer Res 2006; 8: R28. Wienke D, MacFadyen JR, Isacke CM. Identification and characterization of the endocytic transmembrane glycoprotein Endo180 as a novel collagen receptor. Mol Biol Cell 2003; 14: 3592–3604. Clohisy DR, Ramnaraine ML. Osteoclasts are required for bone tumors to grow and destroy bone. J Orthop Res 1998; 16: 660–666. Endo-Munoz L, Cumming A, Sommerville S, et al. Osteosarcoma is characterised by reduced expression of markers of osteoclastogenesis and antigen presentation compared with normal bone. Br J Cancer 2010; 103: 73–81. Clohisy DR, Ogilvie CM, Ramnaraine ML. Tumor osteolysis in osteopetrotic mice. J Orthop Res 1995; 13: 892–897. Nesbitt SA, Horton MA. Trafficking of matrix collagens through bone-resorbing osteoclasts. Science 1997; 276: 266–269. Kato Y, Ozawa S, Miyamoto C, et al. Acidic extracellular microenvironment and cancer. Cancer Cell Int 2013; 13: 89. Dai XM, Zong XH, Akhter MP, et al. Osteoclast deficiency results in disorganized matrix, reduced mineralization, and abnormal osteoblast behavior in developing bone. J Bone Miner Res 2004; 19: 1441–1451. Everts V, Delaisse JM, Korper W, et al. The bone lining cell: its role in cleaning Howship’s lacunae and initiating bone formation. J Bone Miner Res 2002; 17: 77–90. Abdelgawad ME, Soe K, Andersen TL, et al. Does collagen trigger the recruitment of osteoblasts into vacated bone resorption lacunae during bone remodeling? Bone 2014; 67: 181–188. Ohba T, Cole HA, Cates JM, et al. Bisphosphonates inhibit osteosarcoma-mediated osteolysis via attenuation of tumor expression of MCP-1 and RANKL. J Bone Miner Res 2014; 29: 1431–1445. J Pathol 2016; 238: 120–133 www.thejournalofpathology.com


57. Goldsby RE, Fan TM, Villaluna D, et al. Feasibility and dose discovery analysis of zoledronic acid with concurrent chemotherapy in the treatment of newly diagnosed metastatic osteosarcoma: a report from the Children’s Oncology Group. Eur J Cancer 2013; 49: 2384–2391. 58. Fletcher CDMU, Mertens F. World Health Organization Classification of Tumours. Pathology and Genetics of Tumours of Soft Tissue and Bone. IARC Press: Lyon, 2002.

133

59. Jogi A, Pass J, Hoyer-Hansen G, et al. Systemic administration of anti-urokinase plasminogen activator receptor monoclonal antibodies induces hepatic fibrin deposition in tissue-type plasminogen activator deficient mice. J Thromb Haemost 2007; 5: 1936–1944. 60. Jurgensen HJ, Madsen DH, Ingvarsen S, et al. A novel functional role of collagen glycosylation: interaction with the endocytic collagen receptor uparap/ENDO180. J Biol Chem 2011; 286: 32736–32748.

SUPPLEMENTARY MATERIAL ON THE INTERNET The following supplementary material may be found in the online version of this article: Supplementary materials and methods Figure S1 Negative controls for immunohistochemical staining Figure S2 Properties of anti-uPARAP/Endo180 mAb 5f4 in vitro and after administration in vivo Figure S3 Expression of MT1–MMP is unchanged in mAb-treated NCTC-2472 bone tumours Table S1 Osteosarcoma patient material Table S2 Expression of uPARAP/Endo180 and MT1–MMP in bone metastases from breast carcinoma Table S3 Expression of uPARAP/Endo180 in non-malignant osteoblasts/lining cells of osteosarcoma Raw data for microCT-based BMD calculation

