Leading Edge
Previews ER Stress in Dendritic Cells Promotes Cancer Christopher S. Garris1,2 and Mikael J. Pittet1,* 1Center
for Systems Biology, Massachusetts General Hospital Research Institute and Harvard Medical School, Boston, MA 02114, USA Program in Immunology, Harvard Medical School, Boston, MA 02115, USA *Correspondence: [email protected] http://dx.doi.org/10.1016/j.cell.2015.06.006 2Graduate
XBP1 is part of the ER stress response, and when activated in cancer cells, it fosters tumor growth. In this issue of Cell, Cubillos-Ruiz et al. demonstrate that XBP1 in tumor-infiltrating dendritic cells blunts anti-tumor immunity. These findings further imply XBP1 as a relevant target for cancer therapy. Unfolded protein accumulation in the endoplasmic reticulum (ER) activates a cellular stress response that is promoted by the transcription factor XBP1 (Yoshida et al., 2001). In cancer, XBP1 activation within tumor cells directly promotes tumor outgrowth and metastasis (Romero-Ramirez et al., 2004; Chen et al., 2014). XBP1 also regulates immune cell functions, including plasma cell differentiation (Todd et al., 2008) and macrophage proinflammatory cytokine production (Martinon et al., 2010); however, whether or not XBP1 controls tumor-associated immune responses remains unknown. New research by Cubillos-Ruiz et al. in this issue of Cell identifies XBP1 as a critical driver of dendritic cell (DC) dysfunction in the tumor microenvironment. DCs are known to stimulate cytotoxic T cells and can thus promote T cell-mediated tumor rejection. This new study indicates that tumor-associated DCs frequently activate XBP1, which disrupts DC homeostasis and prevents anti-tumor T cell immunity (Cubillos-Ruiz et al., 2015). To investigate ER stress responses within DCs in vivo, Cubillos-Ruiz et al. generated a mouse model that enables selective deletion of XBP1 in DCs (XBP1DC-KO) (Figure 1). By using this model, they showed that XBP1 is not necessary for DC survival. This finding may be unexpected, considering that the ER stress response is usually associated with increased cell survival, but it suggests that ER stress induction in DCs serves functions unrelated to cell preservation. The authors report several notable results from a series of experiments that indicate that XBP1 expression in DCs accelerates primary and metastatic ovarian
carcinoma. For example, they observe that removing XBP1 from DCs at tumor initiation is enough to reduce KrasG12D p53–/– driven ovarian cancer growth. Additionally, mice lacking XBP1 in DCs survive longer when challenged with orthotopic tumors than those with XBP1. Finally, implanting tumor cells admixed with XBP1-sufficient DCs accelerates cancer progression, whereas XBP1-deficient DCs delay this process. Tumor (or tumor stroma)-derived factors likely trigger the ER stress response in DCs because XBP1 activation is significantly higher in tumor-infiltrating DCs than in splenic DCs. Nevertheless, these factors’ identity requires more study. Hypoxia activates XBP1 in cancer cells but did not in DCs, at least in the experimental models used, nor did immunesuppressive cytokines like IL-10 and TGF-b, thereby suggesting that different mechanisms activate XBP1 in cancer cells and DCs. Interestingly, the authors found that the ER stress response in DCs is associated with increased reactive oxygen species (ROS) and lipid peroxidation byproducts. One of these byproducts is the reactive aldehyde 4-hydroxynonenal (4-HNE), which forms stable adducts with numerous proteins. DC treatment with 4-HNE alone was sufficient to activate XBP1 and chaperone protein expression; thus, 4-HNE induces the ER stress response in DCs. 4-HNE was not detected in significant levels in tumor ascites fluid, indicating that it is produced within DCs. Cubillos-Ruiz et al. also report that tumor-infiltrating wild-type DCs accumulate intracellular lipids, whereas XBP1deficient DCs do not. XBP1 deficiency
1492 Cell 161, June 18, 2015 ª2015 Elsevier Inc.
also reduces the expression of multiple genes involved in lipid biosynthetic pathways. Accordingly, XBP1 likely promotes lipid buildup in wild-type DCs (Figure 1). Such lipid accumulation is known to inhibit T cell priming functions (Herber et al., 2010). Consequently, CubillosRuiz et al. report that XBP1 deletion enables surface expression of MHC-I/ peptide complexes in DCs. XBP1 deletion in DCs also increases anti-tumor CD4+ and CD8+ T cells but reduces T regulatory cells in the tumor stroma, presumably due to enhanced antigen presentation. T cells are responsible for inhibiting tumor growth in this experimental context because adoptive T cell transfer from tumor-bearing XBP1DC-KO mice (but not from control animals) reduced tumor burden in recipients challenged with ovarian cancer cells. The anti-tumor effects of XBP1-deficient DCs indicate that targeting XBP1 may have therapeutic value. CubillosRuiz et al. used XBP1-targeting small interfering RNA nanoparticles (referred to as siXBP1) that preferentially deliver their cargo in phagocytes, including DCs and macrophages, in vivo. Administering siXBP1 to ovarian cancer-bearing animals enhanced anti-tumor T cell responses and significantly prolonged mouse survival. siXBP1 treatment failed in T-celldeficient Rag2–/– mice, further indicating that adaptive immune cells mediate the therapeutic response. In sum, this work has multiple levels of significance. The researchers reveal a previously unrecognized mechanism of ER stress induction and lipid accumulation in DCs. The process that they describe inhibits anti-tumor T cell immunity but can be targeted in therapy. In
et al., 2008). Yet the authors also observed that DC treatment with oleate, which may cause lipid accumulation without activating ER stress (Karaskov et al., 2006), also impairs T cell activation. The answers to these questions are important because they should clarify how, and which, DCs are susceptible to tumor co-option. A full understanding of ER stress disruption of DC homeostasis will enable additional avenues of therapeutic discovery.
REFERENCES Broz, M.L., Binnewies, M., Boldajipour, B., Nelson, A.E., Pollack, J.L., Erle, D.J., Barczak, A., Rosenblum, M.D., Daud, A., Barber, D.L., et al. (2014). Cancer Cell 26, 638–652. Chen, X., Iliopoulos, D., Zhang, Q., Tang, Q., Greenblatt, M.B., Hatziapostolou, M., Lim, E., Tam, W.L., Ni, M., Chen, Y., et al. (2014). Nature 508, 103–107.
Figure 1. Model of XBP1-Driven ER Stress in Dendritic Cells (Left) XBP1WT DC. (Right) XBP1KO DC. Tumor-derived factors activate lipid peroxidation within DCs in a ROS-dependent mechanism. Lipid peroxidation byproducts trigger ER stress responses that activate XBP1. XBP1 activation is connected to lipid synthesis pathways that ultimately increase lipid load within the DCs and inhibit DC-mediated T cell priming. Loss of XBP1 eliminates ER stress responses and lipid synthesis/accumulation and enhances DC ability to prime anti-tumor T cell responses.
fact, interventional therapy targeting XBP1 could fight cancers in two ways: by preventing cancer cell survival (when the drugs reach tumor cells) and by fostering protective immunity (when the drugs reach DCs in the tumor microenvironment). In considering how the authors’ results may someday translate to new therapeutic options, some key questions remain unanswered. For example, the factors instigating ROS production and lipid accumulation within DCs require identification. Candidates include TLR ligands because they can activate XBP1 in macrophages (Martinon et al., 2010). Also, it
is unknown whether DCs’ lipid content and ER stress profile increase as the disease progresses or vary among DC subsets. Indeed, tumors contain distinct DC populations that have different capacities to stimulate T cell functions locally (Broz et al., 2014), but whether these DC populations have different responses to ER stressors requires study. Finally, it would be helpful to define whether lipid-mediated inhibition of DC functions always depends on activating an ER stress response. XBP1 regulates lipid metabolism in the liver, and transcriptional signatures indicate that lipid biosynthesis is similarly affected by XBP1 in DCs (Lee
Cubillos-Ruiz, J.R., Silberman, P.C., Rutkowski, M.R., Chopra, S., Perales-Puchalt, A., Song, M., Zhang, S., Bettigole, S.E., Gupta, D., Holcomb, K., et al. (2015). Cell 161, this issue, 1527–1538. Herber, D.L., Cao, W., Nefedova, Y., Novitskiy, S.V., Nagaraj, S., Tyurin, V.A., Corzo, A., Cho, H.-I., Celis, E., Lennox, B., et al. (2010). Nat. Med. 16, 880. Karaskov, E., Scott, C., Zhang, L., Teodoro, T., Ravazzola, M., and Volchuk, A. (2006). Endocrinology 147, 3398–3407. Lee, A.-H., Scapa, E.F., Cohen, D.E., and Glimcher, L.H. (2008). Science 320, 1492–1496. Martinon, F., Chen, X., Lee, A.-H., and Glimcher, L.H. (2010). Nat. Immunol. 11, 411–418. Romero-Ramirez, L., Cao, H., Nelson, D., Hammond, E., Lee, A.-H., Yoshida, H., Mori, K., Glimcher, L.H., Denko, N.C., Giaccia, A.J., et al. (2004). Cancer Res. 64, 5943–5947. Todd, D.J., Lee, A.-H., and Glimcher, L.H. (2008). Nat. Rev. Immunol. 8, 663–674. Yoshida, H., Matsui, T., Yamamoto, A., Okada, T., and Mori, K. (2001). Cell 107, 881–891.
Cell 161, June 18, 2015 ª2015 Elsevier Inc. 1493
Previews ER Stress in Dendritic Cells Promotes Cancer Christopher S. Garris1,2 and Mikael J. Pittet1,* 1Center
for Systems Biology, Massachusetts General Hospital Research Institute and Harvard Medical School, Boston, MA 02114, USA Program in Immunology, Harvard Medical School, Boston, MA 02115, USA *Correspondence: [email protected] http://dx.doi.org/10.1016/j.cell.2015.06.006 2Graduate
XBP1 is part of the ER stress response, and when activated in cancer cells, it fosters tumor growth. In this issue of Cell, Cubillos-Ruiz et al. demonstrate that XBP1 in tumor-infiltrating dendritic cells blunts anti-tumor immunity. These findings further imply XBP1 as a relevant target for cancer therapy. Unfolded protein accumulation in the endoplasmic reticulum (ER) activates a cellular stress response that is promoted by the transcription factor XBP1 (Yoshida et al., 2001). In cancer, XBP1 activation within tumor cells directly promotes tumor outgrowth and metastasis (Romero-Ramirez et al., 2004; Chen et al., 2014). XBP1 also regulates immune cell functions, including plasma cell differentiation (Todd et al., 2008) and macrophage proinflammatory cytokine production (Martinon et al., 2010); however, whether or not XBP1 controls tumor-associated immune responses remains unknown. New research by Cubillos-Ruiz et al. in this issue of Cell identifies XBP1 as a critical driver of dendritic cell (DC) dysfunction in the tumor microenvironment. DCs are known to stimulate cytotoxic T cells and can thus promote T cell-mediated tumor rejection. This new study indicates that tumor-associated DCs frequently activate XBP1, which disrupts DC homeostasis and prevents anti-tumor T cell immunity (Cubillos-Ruiz et al., 2015). To investigate ER stress responses within DCs in vivo, Cubillos-Ruiz et al. generated a mouse model that enables selective deletion of XBP1 in DCs (XBP1DC-KO) (Figure 1). By using this model, they showed that XBP1 is not necessary for DC survival. This finding may be unexpected, considering that the ER stress response is usually associated with increased cell survival, but it suggests that ER stress induction in DCs serves functions unrelated to cell preservation. The authors report several notable results from a series of experiments that indicate that XBP1 expression in DCs accelerates primary and metastatic ovarian
carcinoma. For example, they observe that removing XBP1 from DCs at tumor initiation is enough to reduce KrasG12D p53–/– driven ovarian cancer growth. Additionally, mice lacking XBP1 in DCs survive longer when challenged with orthotopic tumors than those with XBP1. Finally, implanting tumor cells admixed with XBP1-sufficient DCs accelerates cancer progression, whereas XBP1-deficient DCs delay this process. Tumor (or tumor stroma)-derived factors likely trigger the ER stress response in DCs because XBP1 activation is significantly higher in tumor-infiltrating DCs than in splenic DCs. Nevertheless, these factors’ identity requires more study. Hypoxia activates XBP1 in cancer cells but did not in DCs, at least in the experimental models used, nor did immunesuppressive cytokines like IL-10 and TGF-b, thereby suggesting that different mechanisms activate XBP1 in cancer cells and DCs. Interestingly, the authors found that the ER stress response in DCs is associated with increased reactive oxygen species (ROS) and lipid peroxidation byproducts. One of these byproducts is the reactive aldehyde 4-hydroxynonenal (4-HNE), which forms stable adducts with numerous proteins. DC treatment with 4-HNE alone was sufficient to activate XBP1 and chaperone protein expression; thus, 4-HNE induces the ER stress response in DCs. 4-HNE was not detected in significant levels in tumor ascites fluid, indicating that it is produced within DCs. Cubillos-Ruiz et al. also report that tumor-infiltrating wild-type DCs accumulate intracellular lipids, whereas XBP1deficient DCs do not. XBP1 deficiency
1492 Cell 161, June 18, 2015 ª2015 Elsevier Inc.
also reduces the expression of multiple genes involved in lipid biosynthetic pathways. Accordingly, XBP1 likely promotes lipid buildup in wild-type DCs (Figure 1). Such lipid accumulation is known to inhibit T cell priming functions (Herber et al., 2010). Consequently, CubillosRuiz et al. report that XBP1 deletion enables surface expression of MHC-I/ peptide complexes in DCs. XBP1 deletion in DCs also increases anti-tumor CD4+ and CD8+ T cells but reduces T regulatory cells in the tumor stroma, presumably due to enhanced antigen presentation. T cells are responsible for inhibiting tumor growth in this experimental context because adoptive T cell transfer from tumor-bearing XBP1DC-KO mice (but not from control animals) reduced tumor burden in recipients challenged with ovarian cancer cells. The anti-tumor effects of XBP1-deficient DCs indicate that targeting XBP1 may have therapeutic value. CubillosRuiz et al. used XBP1-targeting small interfering RNA nanoparticles (referred to as siXBP1) that preferentially deliver their cargo in phagocytes, including DCs and macrophages, in vivo. Administering siXBP1 to ovarian cancer-bearing animals enhanced anti-tumor T cell responses and significantly prolonged mouse survival. siXBP1 treatment failed in T-celldeficient Rag2–/– mice, further indicating that adaptive immune cells mediate the therapeutic response. In sum, this work has multiple levels of significance. The researchers reveal a previously unrecognized mechanism of ER stress induction and lipid accumulation in DCs. The process that they describe inhibits anti-tumor T cell immunity but can be targeted in therapy. In
et al., 2008). Yet the authors also observed that DC treatment with oleate, which may cause lipid accumulation without activating ER stress (Karaskov et al., 2006), also impairs T cell activation. The answers to these questions are important because they should clarify how, and which, DCs are susceptible to tumor co-option. A full understanding of ER stress disruption of DC homeostasis will enable additional avenues of therapeutic discovery.
REFERENCES Broz, M.L., Binnewies, M., Boldajipour, B., Nelson, A.E., Pollack, J.L., Erle, D.J., Barczak, A., Rosenblum, M.D., Daud, A., Barber, D.L., et al. (2014). Cancer Cell 26, 638–652. Chen, X., Iliopoulos, D., Zhang, Q., Tang, Q., Greenblatt, M.B., Hatziapostolou, M., Lim, E., Tam, W.L., Ni, M., Chen, Y., et al. (2014). Nature 508, 103–107.
Figure 1. Model of XBP1-Driven ER Stress in Dendritic Cells (Left) XBP1WT DC. (Right) XBP1KO DC. Tumor-derived factors activate lipid peroxidation within DCs in a ROS-dependent mechanism. Lipid peroxidation byproducts trigger ER stress responses that activate XBP1. XBP1 activation is connected to lipid synthesis pathways that ultimately increase lipid load within the DCs and inhibit DC-mediated T cell priming. Loss of XBP1 eliminates ER stress responses and lipid synthesis/accumulation and enhances DC ability to prime anti-tumor T cell responses.
fact, interventional therapy targeting XBP1 could fight cancers in two ways: by preventing cancer cell survival (when the drugs reach tumor cells) and by fostering protective immunity (when the drugs reach DCs in the tumor microenvironment). In considering how the authors’ results may someday translate to new therapeutic options, some key questions remain unanswered. For example, the factors instigating ROS production and lipid accumulation within DCs require identification. Candidates include TLR ligands because they can activate XBP1 in macrophages (Martinon et al., 2010). Also, it
is unknown whether DCs’ lipid content and ER stress profile increase as the disease progresses or vary among DC subsets. Indeed, tumors contain distinct DC populations that have different capacities to stimulate T cell functions locally (Broz et al., 2014), but whether these DC populations have different responses to ER stressors requires study. Finally, it would be helpful to define whether lipid-mediated inhibition of DC functions always depends on activating an ER stress response. XBP1 regulates lipid metabolism in the liver, and transcriptional signatures indicate that lipid biosynthesis is similarly affected by XBP1 in DCs (Lee
Cubillos-Ruiz, J.R., Silberman, P.C., Rutkowski, M.R., Chopra, S., Perales-Puchalt, A., Song, M., Zhang, S., Bettigole, S.E., Gupta, D., Holcomb, K., et al. (2015). Cell 161, this issue, 1527–1538. Herber, D.L., Cao, W., Nefedova, Y., Novitskiy, S.V., Nagaraj, S., Tyurin, V.A., Corzo, A., Cho, H.-I., Celis, E., Lennox, B., et al. (2010). Nat. Med. 16, 880. Karaskov, E., Scott, C., Zhang, L., Teodoro, T., Ravazzola, M., and Volchuk, A. (2006). Endocrinology 147, 3398–3407. Lee, A.-H., Scapa, E.F., Cohen, D.E., and Glimcher, L.H. (2008). Science 320, 1492–1496. Martinon, F., Chen, X., Lee, A.-H., and Glimcher, L.H. (2010). Nat. Immunol. 11, 411–418. Romero-Ramirez, L., Cao, H., Nelson, D., Hammond, E., Lee, A.-H., Yoshida, H., Mori, K., Glimcher, L.H., Denko, N.C., Giaccia, A.J., et al. (2004). Cancer Res. 64, 5943–5947. Todd, D.J., Lee, A.-H., and Glimcher, L.H. (2008). Nat. Rev. Immunol. 8, 663–674. Yoshida, H., Matsui, T., Yamamoto, A., Okada, T., and Mori, K. (2001). Cell 107, 881–891.
Cell 161, June 18, 2015 ª2015 Elsevier Inc. 1493
