One reason ovarian cancer is so deadly: it turns off immune cells that try to fight it. A Weill Cornell Medical College team has found that disarming a gene called XBP1 rearms immune cells—which successfully combat ovarian cancer.
In Cell, the team reported that ovarian cancer, in later, liquified stages, oozes about inside the body, creating a toxic tumor microenvironment rife with damaging, reactive oxygen molecules. These molecules prompt the endoplasmic reticulum stress response pathway to turn on the XBP1 gene in dendritic cells (DCs). This prompts lipids to build up inside DCs and render them unable to do their job, which is to engulf cancer bits (antigen) and show them to T cells for attack.
The team is creating drugs to block XBP1 in both cancer and dendritic cells. For the paper, they injected nanoparticles, carrying a genetic molecule (siRNA) that blocks XBP1, into an ovarian cancer mouse model. DCs, seeing the nanoparticles as foreigners, ingested them. Once inside the DCs, like soldiers in a Trojan Horse, the nanoparticles swarmed out, and turned XBP1 off. This freed the dendritic cells to get to work warning T cells about the cancer.
“Probably the most important and clinically relevant finding in the paper is a demonstrated opportunity to therapeutically silence XBP1 in DCs to improve the development of antitumor immune responses,” Michael Shurin, M.D., Ph.D. (ABMLI) told Bioscience Technology. Shurin, University of Pittsburgh Medical Center Clinical Immunopathology associate director, was uninvolved in the work.
“Using nanodelivery of specific siRNA to silence the XBP1 gene of DCs in vivo, the authors demonstrated a significant reduction of tumor burden and tumor progression in mice, mediated by increased activity of tumor-specific cytotoxic T cells. These data provide a new target in key antigen-presenting cells in the tumor milieu to develop a therapeutic approach to ovarian cancer,” Shurin said.
Noted Louisiana State University tumor immunologist Paulo Rodriguez, Ph.D., to Bioscience Technology: “Interestingly, deletion of XBP1 resulted in significant anti-tumor effects, and a dramatic transformation of tumor-associated DCs, from cells that promoted tumor growth, into populations that induced anti-tumor responses through activation of T cells.” Rodriguez was also uninvolved in the study. “These results emphasize the relevance of tumor-induced stress signaling on suppression of protective immune responses in cancer, a field that has generated significant attention over the last years.”
Said ImmunOvar Research Group oncologists An Coosemans and Thais Baert, also uninvolved, to Bioscience Technology: “It is a highly qualitative paper which opens new perspectives in the treatment of ovarian cancer.”
Dendritic cells: big players
Unlike some other cancers, important yet impaired DCs are often found buried and paralyzed in ovarian cancer tumors, like sleeping—or turncoat—generals. This matters, as DCs orchestrate all immune attacks—by, as noted, scooping up bits of foreign cells or cancers and presenting them to killer T cells.
It has been known the endoplasmic reticulum (ER) stress response, and the activation of XBP1, cause tumorigenesis in ovarian cancer. (The healthy ER is an organelle, in all cells, involved in protein and lipid production.) But when the Cornell team, by turns, constitutively over-expressed XBP1 in DCs, then knocked out or inactivated it in DCs, they were the first to discover XBP1 also promotes tumor growth by impairing DCs. Disabling it restores DC function, wipes out tumors, and promotes increased longevity in mice.
Jose Conejo-Garcia of the Wistar Institute, a co-author on the new paper, told Bioscience Technology: “DCs are big players. They play a major role in the biology of ovarian cancer, and interventions that change their phenotype have very significant effects on tumor progression, without any direct intervention on tumor cells. I do not know other tumors where bona fide DCs turned into immunosuppressive, tumor-promoting cells, and play the major role they do in ovarian cancer. Other tumors accumulate other myeloid leukocytes, typically macrophages. But solid tumor masses in ovarian cancer are very rich in tolerogenic dendritic cells for reasons we do not fully understand.”
Conejo-Garcia said his team’s paper “unveils a major mechanism whereby protective immunity against ovarian cancer is abrogated through the activation of the ER stress response in DCs in the tumor microenvironment, which activates a transcriptional program driven by the transcription factor XBP1. XBP1 activation results in the accumulation of lipids, which abrogates the capacity of DCs to effectively present antigen to tumor-reactive T cells.”
As noted, Conejo-Garcia said, leukocytes expressing DC markers “have been previously found to universally accumulate in human ovarian cancer masses, as well as solid tumors and ascites in preclinical mouse models. In many tumors, these inflammatory— but bona fide—DCs outnumber canonical macrophages, which are the predominant leukocyte subset in other epithelial malignancies. Rather than priming protective immunity, ovarian cancer-infiltrating DCs have been associated with immunosuppressive, tumor-promoting activities. The mechanisms driving their aberrant phenotype remained elusive.”
The new paper identifies “what could be the major cause for the abrogation of the capacity of antigen-presenting cells (DCs) to support anti-tumor immunity, at least in ovarian cancer,” he said. “Because silencing XBP1 specifically in DCs at tumor beds enhances protective immunity and leads to tumor regression, this study opens new avenues for therapeutic targeting of ER stress responses through small molecules. These interventions could influence not only tumor DCs, but also the tumor cell itself, which also relies on ER stress to fuel malignant progression. These therapies could synergize with clinically available checkpoint inhibitors and/or CAR T cells, which are particularly promising against ovarian cancer due to its immunogenicity and the paucity of therapeutic alternatives.”
Conejo-Garcia added that DCs infiltrating ovarian tumors exhibit phenotypic attributes of inflammatory DCs, and therefore represent cells that do not exist in the steady state.
“However, they can acquire the capacity to boost anti-tumor immunity through TLR (toll like receptor) activation and, more effectively, inhibition of ER stress. Promoting the immunostimulatory activity of DCs already infiltrating tumor beds could be therefore significantly more effective than boosting tumor-specific responses with exogenous antigen-presenting cells. However, it is possible that lipid accumulation driven by ER stress protects DCs from dying in the tumor microenvironment. New studies should determine the full spectrum of consequences of inhibiting the ER stress response in ovarian cancer DCs, as well as the pathways driving their transformation from a potentially immunostimulatory type, into an immunosuppressive, tumor-promoting type.”
The paper scored a few firsts, Shurin told Bioscience Technology. “Molecular signaling initiated by ER stress in different cell types is a widely studied area of modern oncology,” he said. “The fact that the authors demonstrated a robust expression of total and spliced XBP1 mRNA in tumor-associated DCs is an original finding that may suggest that XBP1 drives tumor growth by inhibiting DC-dependent anti-cancer immunity. This suggestion was proven in the paper. Very important is the fact that XBP1 expression in DCs is crucial for the rapid initiation, and progression, of primary ovarian cancer in mice. And DCs devoid of XBP1 in the tumor milieu display an increased ability to control tumor growth.”
It was known that aberrant lipid accumulation in tumor-associated DCs obstructs their normal antigen-presenting capacity, Shurin told Bioscience Technology. But “molecular mechanisms linking lipid oxidative metabolism with ER stress markers and antitumor immunity are unknown. The discovery that unsaturated aldehyde 4-hydroxy-trans-2-nonenal (4-HNE)—generated by intracellular lipid oxidation (caused) by reactive oxygen species in DCs—is responsible for promoting ER stress and constitutive XBP1 activation….that is the study’s main molecular finding. Given the strong association between excessive lipid accumulation and DC dysfunction in cancer, the proof that XBP1 is a potential mediator of this process is quite important.”
Another key finding, said Shurin, is the paper’s confirmation of animal data in human tumors. “The authors showed that expression of ER stress response markers in tumor-infiltrating DCs was associated with reduced T cell infiltration in human ovarian cancer specimens, which suggests that ER-stressed DCs in the tumor microenvironment may regulate the antitumor activity of T cells.”
And as noted, the most clinically relevant “first,” Shurin told Bioscience Technology, was the silencing of XBP1 to improve immune response.
Rodriguez told Bioscience Technology the new paper successfully defines the role of XBP1 in the suppressive function of tumor-associated DCs. “The results indicate a robust induction of ER stress and XBP1 expression in tumor-associated DCs, induced by endogenous lipid peroxidation-linked processes. The critical effects of XBP1 deletion on the ability of DCs to present antigens, and to support T cell responses, were associated with a decreased accumulation of intra-cellular lipids and impaired triglyceride biosynthesis. In contrast, deletion of XBP1 did not alter the expression of MHC I-II or PDL1 molecules.”
The latter helps explain why DCs survive constitutive XBP1 expression, so key to successfully switching their phenotype from tolerogenic, back to immunogenic. “Importantly, the authors tested the therapeutic effect of silencing of XBP1 in tumor-associated DCs through the use of polyethylenimine nanoparticle encapsulated anti-XBP1 siRNA, which induced significant anti-tumor effects mediated by effector immune responses.”
The results emphasize, Rodriguez said, the key role of tumor-induced stress signaling on reversal of natural immune responses against cancer, which has been much studied of late. “It is generating significant attention. In particular, findings support the described metabolic polarization of DCs in tumor-bearing hosts, and under chronic inflammation, into cells preferentially having metabolism of lipids.”
Shurin told Bioscience Technology this was “an experimental research paper and many additional studies are required. One interesting notion may be use of nanomaterials for modulating cellular functions in vivo: recently, several publications revealed that different nanoparticles in vivo modulate activity of DCs and other myeloid cells, and induce development of allergic reactions and tumor growth. Careful evaluation of nanotoxicities of individual nanomaterials will be important in developing novel nanomedical applications.”
Coosemans and Baert told Bioscience Technology: “Especially in the emerging field of immune checkpoint blockades, this is a fresh, totally new finding. We feel, however, that somewhat lacking is the clinical relevance and applicability. Are there chemotherapeutics or other agents that act against this XBP1? Would there be side effects if one could block XBP1 in humans?”
In the future, Shurin said the biological and clinical significance of XBP1 transcription factor regulation of DC function in cancer “should be determined in different animal tumor models, and tumor specimens obtained from patients with different types of cancer. Another important direction is to develop technologies allowing cell-specific targeting of ER stress components in tumor-associated DCs in vivo. Finally, it will be important to determine tumor-derived factors responsible for modulating activity of tumor-associated DC by triggering ER stress.”
Rodriguez told Bioscience Technology the list of questions that need to be answered is long—but worth the effort.
Asked Rodriguez:“Are lipid accumulation and peroxidation the initial drivers of ER stress in tumor -associated DCs, or a positive feedback from an existing increased and unbalanced ER stress? What are the molecular and metabolic mechanisms by which the ER stress-driven accumulation of lipids in myeloid populations results in the inhibition of T cell responses and promotion of chronic inflammation? Is the induction of ER stress and XBP1 in tumors a specific characteristic of tumor-associated DCs, or is it shared by other myeloid populations, including tumor-associated macrophages, myeloid-derived suppressive cells, tumor-infiltrating neutrophils, among others?”
Other questions, Rodriguez told Bioscience Technology, include: “What are the effects of other mediators of ER stress such as ATF6, IRE1, and PERK in the tolerogenic effect induced by tumor-associated DCs?? Is ER stress in tumor-associated DCs always associated with suppression, or does it depend on tumor burden ? Is the induction of ER and XBP1 in DCs a physiological mechanism to maintain T cell tolerance when needed, i.e. pregnancy?”
Rodriguez concluded: “The investigation of these key questions and others will let the field understand the insight pathways regulating a significant mechanism of evasion of immune responses in tumors.”
Cynthia Fox, Science Writer