Bone-breaking cancer treatment G David Roodman A drug used to counteract low white blood cell counts in individuals with breast cancer may also be inducing bone metastases (pages 62–69). Patients with a variety of tumors, including those with breast cancer, are often treated with granulocyte–monocyte colony stimulating factor (GM-CSF), a cytokine that increases white cell counts. GM-CSF stimulates the proliferation and differentiation of hematopoietic precursors, thereby replenishing blood cells ravaged by chemotherapy1. In this issue, Park et al.2 raise cautions about using GM-CSF in patients with breast cancer. In animal models, they show that GM-CSF promotes breast cancer metastases that destroy bone—a complication that occurs in up to 65–75% of patients3. Bone metastases in individuals with breast cancer predominantly cause bone destruction and can result in severe bone pain, fractures and elevated blood calcium levels, because of the increased release of calcium from the involved bones3–5. Patients with breast cancer are highly susceptible to bone metastasis because the cancer cells express factors that both increase their predilection to go to bone and induce the formation of osteoclasts, bone-resorbing cells. Studies in animal models and with human breast cancer cell lines have helped to identify factors produced by breast cancer cells responsible for their metastasizing to bone. These studies have shown that breast cancer cells produce parathyroid hormone–related protein (PTHrP) and interleukin (IL)-6, which directly or indirectly increase the formation of osteoclasts6. As osteoclasts destroy bone, immobilized growth factors in the bone matrix are released, which then stimulate the growth of the tumor4. The author is in the Center for Bone Biology at the University of Pittsburgh Medical Center and the Myeloma Program of the University of Pittsburgh Cancer Institute, VA Pittsburgh Healthcare System, R & D 151U, University Drive C, Pittsburgh, Pennsylvania 15240, USA. e-mail:
[email protected]
Bone cancer cells in bone metastasis
GM-CSF PTHrP
Growth factors, Ca2+ Osteoclast precursors Osteoblast
Osteoclast
RANKL OPG Bone resorption
Stromal cells
Katie Ris
© 2007 Nature Publishing Group http://www.nature.com/naturemedicine
NEWS AND VIEWS
Bone
Figure 1 Breast cancer cells produce factors such as PTHrP and GM-CSF, which enhance the formation of osteoclasts. GM-CSF expands the osteoclast precursor pool and PTHrP increases RANK ligand and decreases osteoprotegerin (OPG) production by osteoblasts; OPG is a decoy receptor that blocks RANKL. RANKL then induces osteoclast precursor differentiation and increases osteoclast formation. The increase in bone resorption releases growth factors and calcium, which then enhances tumor growth.
Park et al.2 report that human breast cancer cells that are metastatic to bone have increased nuclear factor–κB (NF-κB) signaling that results in the production of high levels of GM-CSF. GM-CSF in turn induces osteoclast formation and bone destruction. To back up this observation, the authors blocked GM-CSF production using a shRNA, or inhibited GM-CSF activity with two different neutralizing antibodies to GM-CSF. Both treatments decreased growth and osteoclastic bone destruction in an animal model of breast cancer used extensively to study the pathogenesis of breast cancer bone metastasis. Other NF-κB–regulated genes that induce osteoclast formation, such as genes encoding IL-8 or IL-6, were also expressed by the breast
NATURE MEDICINE VOLUME 13 | NUMBER 1 | JANUARY 2007
cancer cells—although their expression was not increased to the same extent as GM-CSF. Since GM-CSF can increase osteoclast formation, the authors asked whether GM-CSF expression might be driving metastasis. They confirmed that GM-CSF induced osteoclast formation in human bone marrow cultures and that transfecting breast cancer cells with a super-repressor of NF-κB decreased the capacity of these breast cancer cells to metastasize to bone. Importantly, they showed that transfecting breast cancer cell lines that normally do not metastasize to bone with a GM-CSF expression vector increased their bonemetastatic potential. The finding that GM-CSF increased osteoclast formation in this mouse model of breast
25
© 2007 Nature Publishing Group http://www.nature.com/naturemedicine
NEWS AND VIEWS cancer bone metastasis is surprising, as other researchers have found that GM-CSF is an inhibitor of mouse osteoclast formation7. Studies in human marrow cultures provide some perspective on this conundrum; GM-CSF in these cells increases osteoclast precursor proliferation rather than osteoclast differentiation, thereby expanding the osteoclast precursor pool8,9—but addition of receptor activator of NF-κB ligand (RANKL) or 1,25-dihydroxy-vitamin D3, which induce osteoclast precursor differentiation, is required to increase the number of osteoclasts in these human cells. These results suggest that other factors also produced by breast cancer cells, such as RANKL, IL-8 or PTHrP, are providing the differentiation stimulus required for GM-CSF to increase osteoclast formation and bone metastasis. How the results in this study apply to the development of bone metastasis in
breast cancer patients is unclear. GM-CSF was detectable by immunostaining in 75% of bone-metastatic tumor samples from patients, and GM-CSF expression correlated with nuclear localization of NF-κB. However, there were no data on GM-CSF expression or NF-κB localization in the initial tumor samples in the breast from these patients. The question remains whether GM-CSF is only upregulated in the bone metastases or if primary breast cancer cells with high GMCSF levels or NF-κB activity are more likely to metastasize to bone. Studies to determine whether treating mouse models of breast cancer with GMCSF increases bone metastasis should help to determine whether administering GM-CSF to breast cancer patients puts them at risk. If GM-CSF is increased in primary tumors that have a high potential to metastasize to bone, then one could also identify patients at high
risk for developing bone metastasis by measuring GM-CSF production by the primary tumor. The new findings suggest that GM-CSF should be added to the list of factors involved in breast cancer bone metastasis (Fig. 1) and that GM-CSF may represent a new therapeutic target for such metastasis. 1. Chang, D.Z., Lomazow W., Joy Somberg, C., Stan, R. & Perales, M.A. Hematology 9, 207–215 (2004). 2. Park, B.K. et al. Nature Med. 13, 60–67 (2006). 3. Coleman, R.E. Cancer Treat. Rev. 27, 165–176 (2001). 4. Roodman, G.D. N. Engl. J. Med. 350, 1655–1664 (2004). 5. Mundy, G.R. Nat. Rev. Cancer 2, 584–593 (2002). 6. Kozlow, W. & Guise, T.A. J. Mammary Gland Biol. Neoplasia 10, 169–180 (2005). 7. Udagawa, N. et al. J. Exp. Med. 185, 1005–1012 (1997). 8. MacDonald, B.R. et al. J. Bone Miner. Res. 1, 227– 233 (1986). 9. Menaa, C. et al. J. Clin. Invest. 103, 1605–1613 (1999).
IL-23: a master regulator in Crohn disease Markus F Neurath Three studies should shift thinking about the causes of inflammatory bowel disease. It seems that researchers have been focusing on the wrong cytokine as a driving force. An innovative genetic study in people and two other recent studies in mice uncover a key factor in the pathogenesis of inflammatory bowel disease (IBD)1–3. The research puts the spotlight on the pro-inflammatory cytokine interleukin (IL)-23 (refs. 4,5)—prioritizing this molecule and associated signaling pathways as therapeutic targets in IBD and other autoimmune and chronic inflammatory diseases. The chronic inflammatory bowel diseases ulcerative colitis and Crohn disease are common gastrointestinal diseases in both the US and Europe; their incidence is on the rise and they affect as many as 1 in 250 people. Although mortality is low, people with these diseases often suffer from debilitating symptoms, such as abdominal cramping and bloody diarrhea. Current therapies, such as corticosteroids, generally provide only tran-
Markus F. Neurath is in the Laboratory of Immunology, First Medical Clinic, University of Mainz, Langenbeckstrasse 1, 55101 Mainz, Germany. e-mail:
[email protected]
26
sient or marginal relief, so the need for new approaches is great. For many years, research on IBD focused on another pro-inflammatory cytokine, IL-12. IL-12 consists of two subunits, p35 and p40, and, more than ten years ago, neutralizing antibodies to the p40 subunit were found to suppress established chronic intestinal inflammation in an animal model of Crohn disease6. Subsequent studies showed that such antibodies also suppress established gut inflammation in other murine models of IBD, mediated by T cells producing T-helper 1 (TH1) cytokines7. With the discovery of IL-23 in 2000 (ref. 4), it became clear that more was going on. Antibodies to p40, it turns out, also suppress the activity of IL-23, because IL-23 shares the p40 subunit with IL-12. Thus the role of IL-12 may have been overestimated. Indeed, several recent studies knocking out IL-12 and IL-23 subunits in animal models of inflammatory diseases such as multiple sclerosis and rheumatoid arthritis suggest that IL-23, rather than IL-12, drives chronic inflammation. Furthermore, in contrast to IL-12, IL-23 activates a subset of T cells characterized by the production of the cytokine IL-17. This
so-called TH17 T-cell subset expresses the master transcription factor POPγt and mediates chronic inflammatory and autoimmune diseases in animal models8. To identify polymorphisms associated with IBD, Duerr et al.1 scanned the genome for single nucleotide polymorphisms associated with the disease—a powerful genetic approach that is fast becoming more mainstream [p 25]. The authors found that an uncommon coding variant of the gene encoding the IL-23 receptor (IL-23R) confers strong protection against Crohn disease. In contrast, several noncoding variants of the gene encoding the IL-23 receptor were independently associated with disease susceptibility. Although the functional consequences of such variants remain to be determined, there are several possibilities. For instance, loss-offunction mutations of IL-23R may suppress the activation of pathogenic effector T cells in chronic intestinal inflammation. Alternatively, impaired IL-23R signaling may profoundly affect the innate immune response; IL-23 is also expressed on the surface of macrophages and dendritic cells and may thereby control barrier function and immune response
VOLUME 13 | NUMBER 1 | JANUARY 2007 NATURE MEDICINE
