Article Text

Download PDFPDF

Original article
Pharmacological macrophage inhibition decreases metastasis formation in a genetic model of pancreatic cancer
  1. Heidi Griesmann1,2,
  2. Christof Drexel2,
  3. Nada Milosevic2,
  4. Bence Sipos3,
  5. Jonas Rosendahl1,
  6. Thomas M Gress2,
  7. Patrick Michl1,2
  1. 1 Department of Internal Medicine I, Martin Luther University, Halle, Germany
  2. 2 Department of Gastroenterology and Endocrinology, Philipps University, Marburg, Germany
  3. 3 Department of Pathology and Neuropathology, University Hospital Tübingen, Tübingen, Germany
  1. Correspondence to Professor Dr Patrick Michl, Department of Internal Medicine I, Martin Luther University Halle-Wittenberg, Ernst-Grube-Strasse 40, Halle D-06120, Germany; patrick.michl{at}


Objectives Tumour-associated macrophages play an important role in mediating tumour progression. In pancreatic cancer, infiltrating macrophages are known to mediate tumour progression and have been identified in invasive tumours and in early preinvasive pancreatic intraepithelial precursor lesions. We aimed to study the impact of pharmacological macrophage depletion by liposomal clodronate in a genetic mouse model of pancreatic cancer.

Methods KPC mice (LSL-KrasG12D/+;LSL-Trp53R172H/+;Pdx-1-Cre) were treated for 12 weeks with liposomal clodronate or control liposomes. Tumour and metastasis formation as well as alterations in local and circulating immune cells and cytokines were analysed.

Results Treatment with liposomal clodronate effectively reduced CD11b-positive macrophages both in the pancreas and other organs such as liver, lung and spleen. While tumour incidence and growth were only slightly reduced, metastasis formation in the liver and lungs was significantly diminished after macrophage depletion. This antimetastatic effect was independent of the presence of an endogenous primary tumour, since reduced pulmonary colonisation was also detected in clodronate-pretreated mice after tail vein injection of syngeneic pancreatic cancer cell lines. Macrophage inhibition by liposomal clodronate was associated with significantly impaired angiogenesis, reduced circulating vascular endothelial growth factor levels and decreased circulating CD4+CD25+ T cells. These alterations could be confirmed in an independent macrophage depletion model using CD11b-diphtheria toxin receptor mice.

Conclusions Pharmacological depletion of macrophages in a genetic mouse model of pancreatic cancer markedly reduced metastasis formation and is associated with impaired angiogenesis and reduced CD4+CD25+ T cell levels. Pharmacological targeting of infiltrating macrophages represents a promising novel tool for antimetastatic therapeutic approaches.


Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Significance of this study

What is already known on this subject?

  • Pancreatic cancer is a devastating disease associated with a 5-year survival rate of approximately 5% whose resistance to chemotherapy is modulated by an extensive stromal reaction.

  • Tumour-associated macrophages are important constituents of the tumour stroma and are known to mediate tumour progression and metastasis formation.

  • Liposomal clodronate is an effective tool for pharmacological macrophage depletion.

What are the new findings?

  • Application of liposomal clodronate diminishes macrophages in the primary tumour and target organs for metastasis such as liver and lung.

  • Depletion of tumour-associated macrophages results in significantly reduced metastasis formation, which was associated with impaired angiogenesis and altered circulating CD4+CD25+ T cells.

  • The antimetastatic effect most likely represents a systemic effect targeting premetastatic niches in target organs for metastasis such as liver and lung.

How might it impact on clinical practice in the foreseeable future?

  • Pharmacological targeting of macrophages represents a promising antimetastatic therapeutic approach.

  • Combination of bisphosphonates with cytotoxic therapy may open new avenues in the adjuvant and palliative therapy of pancreatic cancer.


Immune cells infiltrating the peritumoural stroma represent a characteristic feature of several solid tumours including pancreatic ductal adenocarcinomas (PDAC). Tumour-associated macrophages (TAMs) constitute a prominent component of the inflammatory stromal infiltration in PDAC.1 TAMs have been recognised as important trigger of therapy resistance and are associated with poor prognosis.2 Soluble factors secreted from infiltrating macrophages have been shown to contribute to cancer progression by mediating resistance to chemotherapeutic drugs, facilitating metastasis and fostering evasion of antitumour immune responses.3–5

Macrophages are characterised by a high functional plasticity and acquire distinct, tissue-specific phenotypes in response to signals present within individual microenvironments.6 In vitro, they can be polarised into opposing functional states, which have been named M1 and M2 polarisations in analogy to the TH1 and TH2 programming of adaptive immune cells. The M1 polarisation can be induced by treatment with lipopolysaccharide (LPS) or tumor necrosis factor alpha (TNF-α), leading to the expression of a proinflammatory cytokine pattern.7 Alternatively, activated M2 macrophages can be polarised by stimulation with interleukin (IL) 4/IL-13, IL-10 or transforming growth factor beta (TGF-β). M2 macrophages may execute tumour-promoting functions by enhancing tissue remodelling, facilitating angiogenesis or regulating immune responses. In most solid tumours including PDAC, TAMs show preferential expression of marker genes associated with an M2 phenotype.8

Based on these data indicating a tumour-promoting role of TAM, several attempts have been made to target the protumoural macrophage action in order to diminish tumour progression and enhance therapeutic efficacy of cytotoxic drugs. In this context, bisphosphonates such as clodronate conjugated with liposomes have been shown to effectively deplete macrophages when administered to mice. After systemic injection, liposomes are ingested by macrophages followed by intracellular release of clodronate and accumulation of clodronate leading to subsequent apoptosis of the macrophage.9

In this study, we evaluated the impact of systemic macrophage depletion on tumour progression and metastasis in a genetic mouse model of pancreatic cancer. Interestingly, we observed a markedly reduced metastasis formation accompanied by substantial changes in angiogenesis and infiltration of other immune cell populations.


Liposomal clodronate effectively diminishes macrophages in vivo

Liposomal clodronate has been frequently used as an effective tool to deplete macrophages in various in vivo systems.10 To study the impact of liposomal clodronate in a genetic mouse model of pancreatic cancer, we used KPC mice expressing activated Kras and inactive p53 under the control of the pancreas-specific promoter Pdx-1.11 These mice develop pancreatic cancers that clinically and histologically recapitulate human disease. We treated the mice with repeated intraperitoneal injections of clodronate liposomes (figure 1A) during the time period from weeks 8 to 20 (figure 1B), during which invasive and metastatic tumours frequently develop from preinvasive pancreatic intraepithelial neoplastic (PanIN) lesions.11 At week 20, mice were sacrificed and tissues as well as blood were subsequently analysed. Efficacy of clodronate liposomes in depletion of myeloid cells was confirmed by flow cytometry (figure 1C, see online supplementary figure S1A,B), on RNA levels (see online supplementary figure S1C) and by immunohistochemistry (figure 1D). Treatment with clodronate liposomes for 2 weeks decreased CD68+ cells in pancreas, spleen or bone marrow of KPC mice by 48%, 56% and 73%, respectively, as assessed by flow cytometry (figure 1C). Since infiltrating myeloid cells can be subdivided into the CD11b+Gr1− cell population, which mainly represents macrophages and monocytes, and CD11b+Gr1+ cells, which represent mainly myeloid-derived suppressor cells (MDSC), we aimed to assess the impact of liposomal clodronate on these two distinct populations. Interestingly, the CD11b+Gr1− cell population representing macrophages and monocytes was significantly reduced upon clodronate treatment (see online supplementary figure S1A). In contrast, CD11b+Gr1+ MDSC were not affected (see online supplementary figure S1B). Upon sacrifice of the mice, immunohistochemical quantification confirmed a significant reduction of CD68+ cells in pancreas, liver and lung, indicating a sustained effect of clodronate liposomes (figure 1D).

Figure 1

Clodronate reduces CD68+ myeloid cells in KPC mice. (A) Model of liposome spheres consisting of phospholipid bilayers encapsulating dichloromethylene bisphosphonate (Cl2MBP; black polygones). (B) Treatment scheme depicting intraperitoneal application of clodronate liposomes or PBS liposomes in KPC mice for 12 weeks (P56-P140). (C) Percentage of CD68+ cells in pancreas, spleen or bone marrow of KPC mice (n=3) after short-term treatment (2 weeks) with clodronate liposomes or PBS liposomes, as determined by flow cytometry. (D) Representative immunohistochemistry (top) of CD68+ cells in sections from pancreas, liver and lung obtained from KPC mice treated with clodronate liposomes (i–iii) or PBS liposomes (iv–vi) for 12 weeks. Original magnification: 400× (pancreas, liver) and 200× (lung). Quantification (below) of CD68+ cells in pancreas, liver or lung from clodronate-liposome (grey symbols; n=20) or PBS-liposome (black symbols; n=21)-treated mice determined by immunohistochemistry. Data shown as mean±SD. Symbols: dots represent data from each animal; black horizontal line, mean percentage for each group. ns; **p=0.0057, ***p=0.0006 by Mann–Whitney U test.

Liposomal clodronate affects metastasis formation

Application of liposomal clodronate did not significantly affect survival of the mice during the time period from weeks 8 to 20 when the experiment was terminated (figure 2A). Likewise, tumour incidence and tumour volume upon sacrifice did not significantly differ between treatment groups (figure 2B; see online supplementary figure S2A). There was a trend towards less advanced PanIN lesions and less dedifferentiated tumours in the clodronate group which, however, did not reach significance due to the limited sample size (see online supplementary figure S2B, C).

Figure 2

Clodronate reduces metastases formation in KPC mice. (A) Survival of KPC mice treated with clodronate liposomes (grey, n=25) or PBS liposomes (black, n=26) from day P56 till P140. Kaplan–Meier curves show a median survival for clodronate of 130 days and for PBS of 117.5 days without significant difference by log-rank test, p>0.05. (B) Analysis of tumour incidence in clodronate-liposome and PBS-liposome mice. ns by Fisher's exact test, p>0.05. (C) Percentage of tumour-bearing KPC mice with metastases after treatment with clodronate liposomes or PBS liposomes. In the clodronate-liposome group (grey bar) 1 of 14 mice and in the PBS-liposome group (black bar) 8 of 17 mice showed metastatic disease. *p≤0.02 by Fisher’'s exact test. (D) Mean number of metastases in livers and lungs of clodronate-liposomes-treated or PBS-liposomes-treated tumour-bearing KPC mice. Below, a contingency table depicts the number and ratio of liver and lung metastases in each treatment group. *p=0.02 by Fisher’'s exact test. M, number of metastases; T, total number of animals per group. (E) Representative H&E staining of metastases. Well-differentiated liver (i) and lung (ii) metastases in a 15-week (P106) old PBS-liposome-treated KPC mouse and well-differentiated liver (iii) metastasis in a 18-week (P130) old clodronate-treated KPC mouse without any lung metastasis (iv). Original magnification: 50× (liver) and 200× (lung). m, metastasis.

In contrast to the minor impact on the primary tumours, the incidence of metastases to liver and lung was significantly reduced after treatment with liposomal clodronate, indicating a pronounced effect of macrophage modulation on metastasis formation (figure 2C–E).

To investigate whether the antimetastatic effect is dependent on the presence of an endogenous primary tumour or rather correlates with systemic alterations in the circulation or in premetastatic niches of target organs, we performed pulmonary metastasis assays with syngeneic KPC tumour cell lines injected in the tail vein of wild-type mice (figure 3A). The mice were pretreated with clodronate liposomes or phosphate-buffered saline (PBS) liposomes for 1 week and received treatment for three additional weeks after tumour cell injection (figure 3A). Interestingly, the size of pulmonary metastases was significantly reduced in mice that had been treated with liposomal clodronate (figure 3B, C), whereas the number of metastases was not significantly different (see online supplementary figure S3). This indicates that liposomal clodronate primarily affects the survival of tumour cells in (pre)metastatic niches of target organs and is independent of the presence of an endogenous primary tumour.

Figure 3

Clodronate reduces metastases formation independently of a primary tumour. (A) Non-transgenic wild-type (WT) mice were treated with clodronate liposomes or PBS liposomes from P56 until P84. Syngeneic KPC cells were administered via tail vein injection at P63 and animals were sacrificed at age of 12 weeks. (B) Quantification of metastasis size in H&E stained lung sections from clodronate-liposome (n=8) and PBS-liposome (n=7)-treated animals. Whole sections were scanned and quantified. Data represent mean area per section±SD. *p=0.018. (C) Representative H&E stained sections of lungs from clodronate-liposome-treated or PBS-liposome-treated animals. Arrows indicate metastases.

M2 polarised macrophages are susceptible to liposomal clodronate

Since M2 polarised macrophages are known to populate advanced tumours and metastases2 and to promote tumour progression,8 we investigated susceptibility of unpolarised, M1 and M2 polarised macrophages to liposomal clodronate treatment in vitro. We could confirm that isolated bone marrow-derived macrophages (BMMs) from non-transgenic mice can be polarised into M1 or M2 phenotypes and express the characteristic marker profiles (see online supplementary figure S4). Interestingly, macrophages polarised towards an M1 phenotype by LPS, interferon gamma (IFNγ) or their combination did not show any susceptibility to clodronate treatment, as assessed by an ATP-based viability assay (figure 4A) and by poly (ADP-ribose) polymerase 1 (PARP) cleavage (figure 4B). In contrast, M2 polarised macrophages were significantly affected by treatment with liposomal clodronate. Likewise, unpolarised macrophages also showed significantly reduced viability after clodronate treatment in vitro (figure 4A). This indicates that M1 macrophages are largely resistant, whereas tumour-associated M2 macrophages are susceptible to clodronate treatment. This susceptibility of M2 macrophages can be attributed to higher uptake of clodronate liposomes based on the higher capability of phagocytosis detected in M2 macrophages when compared with M1 macrophages (see online supplementary figure S5A, B).

Figure 4

M2 polarised macrophages are susceptible to clodronate. (A) Murine bone marrow-derived macrophages from non-transgenic mice were M1 polarised with lipopolysaccharide (LPS), INFγ and LPS/IFNγ or M2 polarised with IL4 and subsequently treated with clodronate liposomes or PBS liposomes for 24 h. Assays were performed in triplicate and cell viability was determined by ATP-based CellTiter Glo assay. Data are presented as mean±SD and are representative of three independent experiments. **p=0.007, ***p=0.002 by Mann–Whitney U test. (B) western blot analysis of PARP expression from clodronate-liposome-treated or PBS-liposome-treated M1 and M2 macrophages (bone marrow-derived macrophages) derived from non-transgenic mice. Actin served as control. fl, whole PARP; cl, PARP-1 fragment at 89 kDa. (C–F) The M1 marker Cxcl10 and M2 markers Ym1, Mrc1 were quantified in lung (C and D) and liver (E and F) tissue of clodronate-liposome-treated or PBS-liposome-treated animals (lung n=4 each; liver n=6 each) by quantitative real-time (qRT)-PCR and were normalised to RPL0. Data are shown as mean±SD. ns; *p≤0.04, ***p=0.002 by Mann–Whitney U test.

To test if this preference for M2 polarised macrophages holds true in vivo, we analysed the expression of a panel of M1 and M2 markers in lung and liver tissues of animals treated with clodronate liposomes or PBS liposomes. As expected, M1 markers such as Cxcl10 were not affected by clodronate (figure 4C, E). In contrast, M2 markers such as Ym1 and Mrc1 were significantly reduced (figure 4D, F), corroborating the increased susceptibility of M2 macrophages for clodronate treatment.

Macrophage depletion is associated with reduced angiogenesis

TAMs are known as important modulators of tumour angiogenesis.6 Based on these data, we investigated the impact of liposomal clodronate on angiogenesis both in the primary tumour and in the target organs. Microvessel density in the pancreatic tumours assessed by CD31 immunohistochemistry was significantly reduced upon liposomal clodronate treatment compared with PBS liposomes (figure 5A, B). Macrophages represent a major source of secreted vascular endothelial growth factor (VEGF) as a paramount mediator of angiogenesis.6 Therefore, we examined local and circulating VEGF levels. We could confirm that the reduction in microvessel density was associated with markedly reduced local and circulating VEGF levels upon treatment with liposomal clodronate, as assessed in liver tissues and mouse sera by western blot using an anti-VEGF antibody (figure 5C, D; for quantification see online supplementary figure S6A, B).

Figure 5

Macrophage depletion reduces angiogenesis, vascular endothelial growth factor (VEGF) expression and T cell infiltration. (A) Representative CD31 immunohistochemistry showing microvessels in pancreatic sections obtained from KPC mice treated with clodronate liposomes (left) or PBS liposomes (right) for 12 weeks. Original magnification 400×. (B) Quantification of CD31+ microvessels in the pancreas of tumour-bearing mice after clodronate-liposome (n=13) or PBS-liposome (n=17) treatment, as determined by immunohistochemistry. Each dot depicts one animal; black horizontal line, the mean for each group. *p=0.0265 using Mann–Whitney U test. (C–F) western blot analysis of VEGF expression in liver tissue lysates (C) and serum (D) samples from clodronate-liposome-treated or PBS-liposome-treated animals, in bone marrow-derived macrophages (E) treated with clodronate liposomes or PBS liposomes for indicated time points and in serum samples from CD11b-DTR mice treated with diphtheria toxin (DT) or PBS. Actin and albumin served as controls. (G–H) DT affects CD4+CD25+ T cell infiltration. CD4+CD25+ T cells from pancreata (G) and blood (H) of DT-treated or PBS-treated CD11b-DTR mice (D–F) were quantified by flow cytometry after gating on live CD45+CD3+ cells. Graphs represent the mean±SD of CD4+ CD25+ T cells among total CD45+CD3+ cells. Dots represent data from each animal; black horizontal lines, the mean percentage for each group. *p<0.04 by Mann–Whitney U test.

To examine whether clodronate liposomes directly affect VEGF levels in macrophages or whether reduced VEGF levels are merely due to reduced macrophage viability, we performed short-term incubations of isolated BMMs with clodronate liposomes or PBS liposomes. VEGF protein levels were reduced as soon as 30 min after addition of liposomal clodronate, consistently up to 120 min (figure 5E; for quantification see online supplementary figure S6C). At these early time points, no impaired viability or signs of apoptosis were detected (not shown), indicating that clodronate liposomes exert a direct short-term impact on VEGF levels in macrophages in addition to their long-term depleting effects.

To rule out that liposomal clodronate affects VEGF levels independently of its effects on macrophages, we employed an independent, diphtheria toxin receptor (DTR)-based mouse model of macrophage depletion (CD11b-DTR) crossed with KPC mice (KPCD).12 In this model, CD11b+ myeloid cells are depleted upon administration of diphtheria toxin (DT) in vivo. Systemic administration of DT in KPCD mice resulted in marked reduction of circulating and local CD11b+Gr1− myeloid cells in pancreas and liver (see online supplementary figure S7A). In contrast, CD11b+Gr1+ MDSC were not reduced, but even appeared slightly increased in the pancreas without being altered in other tissues (see online supplementary figure S7B). Reduction of CD11b+ cells was associated with markedly reduced VEGF levels in the serum of DT-treated animals (figure 5F; for quantification see online supplementary figure S6D). This indicates that alterations in VEGF levels are indeed dependent on altered secretion by CD11b+ myeloid cells.

As described above for clodronate treatment, we also examined potential alterations in M1 and M2 markers in the target tissues liver and lung after DT treatment. None of the examined M1 or M2 markers showed a significant change in expression in the liver or lung following DT treatment, with only the M1 marker Cxcl10 slightly, but non-significantly increased in the liver after DT treatment (see online supplementary figure S8).

Macrophages depletion affects immune cell infiltration

Tumour-associated myeloid cells such as macrophages and MDSC are known to affect attraction and activity of other immune cells such as CD8+ cytotoxic T cells and CD4+CD25+ T cells, which mainly comprise regulatory T cells.13 Therefore, we aimed to investigate the impact of liposomal clodronate on infiltrating T cells in the KPC pancreatic cancer model. Therefore, we isolated T cell populations from different tissues and the systemic circulation from KPC mice treated with liposomal clodronate liposomes or PBS liposomes and KPCD mice treated with DT or PBS. We could show that liposomal clodronate application led to a significant reduction of CD4+CD25+ T cells both locally in the pancreas (figure 5G) and in the systemic circulation (figure 5H) of KPCD mice. The same trend, although not reaching statistical significance, was also seen in KPC mice after application of liposomal clodronate (see online supplementary figure S9A–C). Liposomal clodronate itself had no significant effect on CD4+CD25+ T cells as shown by an ATP-based viability assay after incubation with clodronate liposomes in vitro (see online supplementary figure S9D). This indicates a profound effect of macrophages on CD4+CD25+ T cells, which can be inhibited by using liposomal clodronate or a DT-based genetic model. In contrast to the CD4+CD25+ T cell population, CD4+ or CD8+ T cell populations in the pancreas or the liver were not significantly affected by either clodronate or DT-based macrophage depletion (data not shown). This could also be confirmed by the expression of effector T cell markers in liver tissue of KPC-treated animal (data not shown).


In this study, we identified systemic clodronate application as promising therapeutic avenue for metastasis prevention in a preclinical model of pancreatic cancer. Prolonged administration of liposomal clodronate diminished metastasis formation, which was associated with reduced infiltrating macrophages, reduced angiogenesis and altered immune cell infiltration. These effects were not restricted to liposomal clodronate but were also observed after depletion of macrophages using a second, DT-based genetic mouse model, indicating that the antimetastatic phenotype may generally apply to various means of therapeutic macrophage targeting.

TAMs have been associated with poor prognosis in numerous tumour entities such as breast cancer,14 gastric cancer15 and pancreatic cancer.8 Macrophages may promote tumour progression by enhancing chronic inflammation, matrix remodelling and facilitating tumour cell invasion, intravasation, extravasation and seeding at distant sites.16 ,17 By using pulmonary colonisation assays, we could show that the antimetastatic effect of liposomal clodronate was independent of the presence of an endogenous primary tumour. This suggests that macrophage depletion predominantly affects extravasation, homing and survival of the tumour cells in (pre)metastatic niches of target organs. Interestingly, we found that after tail vein injection, the size of pulmonary metastases was significantly reduced, however, without significant impact on the number of colonies. The tail vein injection model is most likely not suitable to evaluate the homing of single cells derived from endogenous tumours in these niches since multiple cells enter circulation simultaneously after tail vein injection. Our findings are well in line with other reports indicating that macrophages modulate invasiveness and progression of the primary tumour, and populate and prepare premetastatic niches in target organs for survival and persistent growth of metastatic cells.18 Since over 90% of cancer deaths from solid tumours are due to metastasis, the prometastatic function of macrophages represents an intriguing therapeutic approach of high clinical relevance.19

Numerous reports in various cancer models suggest that TAMs affect angiogenesis.19 ,20 TAMs have been identified as a major source of VEGF secretion. VEGF has been demonstrated to be a major contributor to angiogenesis and is indispensable for production of chemokines and additional proangiogenic factors.21 ,22 The expression of VEGF in TAMs is directly regulated by oncogenic transcription factors such as HIF-1a, NF-kB, STAT3 and AP-1.16 Anti-VEGF treatment using neutralising monoclonal antibodies is currently the predominant antiangiogenic therapy for various GI cancers. However, the recruitment of myeloid cells including TAMs into tumours has been shown to curtail the effects of anti-VEGF therapy.23 Intense TAM infiltration as observed in pancreatic cancer might therefore be one of the potential mechanisms for failure of antibody-based anti-VEGF strategies in pancreatic cancer.

In the primary tumour, TAM-dependent upregulation of VEGF results in the so-called angiogenic switch that leads to enhanced microvessel formation and tumour progression. In addition to their accumulation in the primary tumour—and in line with our findings—VEGF-expressing and VEGF receptor 1-expressing myeloid cells including TAMs are also found in premetastatic niches of target organs such as liver and lung.19 The myeloid cells recruited to these sites are known to enhance extravasation of tumour cells leading to increased vascular permeability. Ablation of myeloid cell infiltration, for example, by inhibition of CCR2 signalling, blocks tumour cell extravasation and formation of premetastatic niches, resulting in decreased metastatic seeding.19 ,24 Additional data suggest that targeting tumour-infiltrating macrophages leads also to decreased numbers of tumour-initiating cells, potentially by decreased VEGF production which links angiogenesis and tumour initiation at metastatic sites.25 ,26 In addition to antibody-based or small molecule-based strategies targeting macrophage-derived proteins such as CCR2 or CSF1 that are currently under clinical evaluation, systemic application of macrophage-modulating bisphosphonates therefore represents a promising therapeutic avenue which is, in contrast to various antibody-based strategies, characterised by a highly favourable toxicity profile.

In addition to their proangiogenic effects, TAMs are known as potent immunosuppressors that limit the cytotoxic activity of CD8+ cytotoxic T cells in progressing tumours.6 ,27 In human breast cancer tissues it could be demonstrated that stromal TAM density inversely correlates with CD8+ T cell infiltration.27 The suppression of CD8+ T effector cells is mediated either directly or indirectly by induction and recruitment of regulatory T cells (Treg) to the tumour microenvironment.19 In accordance to these data, we confirmed a marked reduction of Treg cells after application of liposomal clodronate in our preclinical pancreatic cancer model. We were unable to detect a significant upregulation of CD8+ T cells in the primary tumour or in the target organs. Effects on CD8+ T cells were mainly described for CD11b+Gr1+ MDSC.5 This myeloid subpopulation, however, was not significantly affected by clodronate treatment in contrast to the CD11b+Gr1− macrophage population, which was significantly reduced by clodronate. It may be hypothesised that CD11b+Gr1− macrophages exert more prominent effects on tumour angiogenesis and attraction of regulatory T cells while CD11b+Gr1+ MDSC have greater impact on cytotoxic CD8+ T cell inhibition.28

Although our data strongly suggest a causal link between macrophage-derived VEGF, macrophage-dependent modulation of infiltrating regulatory T cells and metastasis, it would be necessary to perform experiments using anti-VEGF strategies together with anti-Treg strategies to formally prove this concept. However, due to the existing body of evidence on the prometastatic effects of both circulating VEGF levels and regulatory T cells, the identified mechanisms are very likely to play a fundamental role as mediators of the observed antimetastatic phenotype.

Clodronate liposomes have been widely used as experimental tool for macrophage depletion in several disease model systems including atherosclerosis, autoimmune disorders and cancer.10 Although the inhibitory effect of liposomal clodronate on macrophages is well established, possible effects on other cell types cannot be ruled out. Antitumour effects of liposomal clodronate have been described in several xenograft tumour models.29–31 In line with our findings, application of liposomal clodronate has been associated with reduced tumour angiogenesis in sarcoma and ovarian cancer xenograft models.30 ,31 However, the inflammatory stroma reaction observed in xenograft models or syngeneic tumour models is known to differ from that in endogenous tumours of genetic animal models. This represents a caveat for the interpretation of xenograft studies and their transferability to human disease. To our knowledge, our study is the first to demonstrate antimetastatic effects of bisphosphonates in endogenous tumours of a genetic mouse model.

Depletion of TAMs with liposomal-encapsulated clodronate was found to act synergistically in combination with other antiangiogenic agents such as sorafenib in a hepatocellular carcinoma (HCC) mouse model,32 suggesting possible synergistic effects between macrophage depletion by bisphosphonates and antiangiogenic-targeted therapies.

Interestingly, the macrophage-depleting effect appears not to be restricted to liposomal-encapsulated clodronate but has also been observed with other bisphosphonates such as zoledronate. In vitro, zoledronate at physiological concentrations reduced viability of macrophages without affecting cancer cells.33 In xenograft models and a genetic breast cancer model in vivo, systemic zoledronate application diminished intratumoural macrophages.34 ,35

In clinical routine, bisphosphonates are used as inhibitors of osteoclasts, a special subgroup of bone macrophages, for treatment of osteoporosis as well as bone metastases in advanced cancers. Interestingly, clinical trials testing adjuvant application of zoledronate in early-stage breast cancer observed a significant risk reduction for the development of bone and according to one randomised trial of visceral metastasis.34 This suggests an antimetastatic effect of zoledronate in addition to its well-known inhibitory effect on osteoclasts. In line with our findings, it might be hypothesised that its effect on the development of visceral metastases might be—at least in part—mediated by inhibitory effects on macrophages in premetastatic niches. In the clinical setting, unbound zoledronate appears also to be effective, although due to its short half-life in circulation, coating with liposomes might increase its half-life and enhance its affinity for cells with capacity for phagocytosis such as macrophages. Generally, the concept of coating cytotoxic drugs such as doxorubicin or cytarabin with liposomes is feasible and current clinical data indicate decreased toxicity and better tolerability.36

In addition to antibody-based or small molecule-based approaches to target macrophages, such as anti-CSF1 antibody or anti-CSF1R small molecule inhibitors,37 which are currently under evaluation in clinical trials, application of bisphosphonates, either liposome-coated or unbound, represents a promising avenue for preventing metastasis and tumour progression. Interestingly, preclinical xenograft studies also suggest a synergistic effect of macrophage inhibition by bisphosphonates and concurrent cytotoxic chemotherapy.34 Based on these data, further clinical trials are warranted to evaluate the role of bisphosphonate in combination with cytotoxic therapy in the adjuvant and palliative therapy of pancreatic cancer or as strategy for patients with high-risk constellations, for example, familial pancreatic cancer.

Materials and methods

Animal models

LSL-KrasG12D/+;LSL-Trp53R172H/+;Pdx-1-Cre (KPC) mice have been previously described.11 C57BL/6 and CD11b-DTR mice12 were obtained from Charles River Laboratories. CD11b-DTR animals expressing the receptor for the human DTR under the control of the CD11b promoter were bred with KPC mice to generate LSL-KrasG12D/+;LSL-Trp53R172H/+;Pdx-1-Cre; CD11b-DTR (KPCD) mice. Eight-week-old (P56) mice were used for all experiments. Blood and tissues were obtained at the indicated time points and processed for PCR analysis, western blot, histology or flow cytometry. For RNA and protein analyses tissue pieces snap frozen were in liquid nitrogen. Blood was allowed to clot at 4°C and serum collected by centrifugation (10.000 rpm, 5 min) was stored in aliquots at −80°C. Mice were maintained in a climate-controlled specific pathogen-free facility. All animal experiments were approved by the local government authorities and performed according to the guidelines of the animal welfare committee.

Macrophage depletion

Clodronate liposomes and PBS liposomes were purchased from Dr N Van Rooijen (Vrije Universiteit, Amsterdam, Netherlands) and prepared as previously described.38 For long-term therapy KPC mice were injected intraperitoneally twice per week with clodronate liposomes (1.4 mg/20 g body weight) or with an equivalent volume of PBS liposomes until week 20 (P140). KPCD mice received one intraperitoneal injection of DT (25 ng/g body weight; List Biological Laboratories) or matched volume of PBS. Monocyte depletion was analysed 24 h later by flow cytometry.



  • Contributors Study concept and design: HG, TMG and PM; acquisition of data: HG, CD, NM and BS; analysis and interpretation of data: HG, CD, NM and BS; drafting of the manuscript: HG and PM; critical revision of the manuscript for important intellectual content: TMG and JR; obtained funding: PM and TMG.

  • Funding This work was supported in part by grants of the European Commisson FP7 grant (Collaborative Project ‘EPC-TM-Net’, to PM and TG), Deutsche Forschungsgemeinschaft (to PM), the LOEWE initiative of the state of Hessen, the Behring-Roentgen Foundation (to PM) and the Deutsche Krebshilfe (to PM). This publication reflects only the authors’ views. The European Community is not liable for any use that may be made of the information herein.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.