Chemotherapy and the Extra-Tumor Immune Microenvironment: EXTRA-TIME

It is now more than 50 years since President Nixon announced his “War on Cancer” in 1971 and more than 7 years since Vice President Biden launched the “Cancer Moonshot” in 2016. More than $3 billion in committed funds have been spent, laying the foundation for several major scientific advances that have expanded treatment options and dramatically improved survival outcomes. However, we are yet to achieve a cure for the most common cancers, including breast, prostate, lung, and colorectal cancer. Globally, an estimated 10 million people continue to die from cancer every year, and based on current cancer trends, this number is predicted to significantly increase over the coming decade (1).

Perhaps the most urgent issue we face is the emergence of treatment-resistant metastatic disease, which today accounts for up to 90% of cancer-related deaths. Despite measures that have significantly improved survival, including the early detection and improved control of primary tumors, the use of combination chemotherapies, radiotherapy, and targeted therapies have made little impact on metastatic relapse outcomes. At the same time, we have also learned that cancer immunotherapy has not only improved survival but also achieved dramatic cure rates in some patients with relapsed/refractory metastatic disease. This therapeutic step-change calls for an overhaul of our thinking. It is time to ask if current treatment strategies that successfully gain control of the primary tumor come at the cost of a greater likelihood of metastatic relapse. What if some treatments, many of which are now standard of care, are doing more harm than good?

To explore this possibility, we believe it is imperative to consider the impact of contemporary therapies outside the tumor on the extra-tumor immune microenvironment (EXTRA-TIME), where maximum tolerated dose (MTD) of chemotherapies may inadvertently disrupt natural antitumor defenses in distal organs and tissues (Fig. 1). This hypothesis is not without merit because accumulating clinical data suggest that some therapies are at best futile and at worse shortening patient survival. On the flip side, the abscopal effect, whereby local irradiation induces a systemic immune response, is a powerful reminder that we will only be able to understand and harness the underlying mechanisms if we go into EXTRA-TIME.

So how do we thoughtfully dose, combine, and sequence treatments with the goal not only to minimize systemic effects on tissues, organs, and host antitumor immunity but also to optimize the ability to generate tumor-specific immunity and cure patients?

Accumulating clinical data suggest that chemotherapy may be altering the course of cancer progression. Patients who present with metastatic disease at the time of primary tumor diagnosis respond better to therapy than patients who relapse with metastatic disease at a later date (2). The pattern of metastatic disease is changing. In an analysis of 60,277 patients diagnosed with early breast cancer from 1978 to 2013, the rate of lung metastases remained stable while the rate of bone metastases reduced by 50%; however, the percentage of liver and cerebral metastases more than doubled and was associated with decreased postmetastatic survival (3).

The nature of metastatic disease is also changing. Analysis of 6,361 women with invasive breast cancer who underwent surgery over 4 decades from 1975 to 2006 at Guy's Hospital, London, has shown that the likelihood of developing metastasis and overall mortality has steadily declined (4). However, the metastasis-free interval shortened and postmetastatic survival decreased by 6.6 months, indicating an aggressive treatment-resistant subpopulation of cancer cells may be causing early death in a subgroup of women. Could these result from time-related changes in the use and impact of different types of adjuvant chemotherapy, such as taxanes and anthracyclines? Investigators have suggested this represents an “adjuvant therapy-related shortening of survival” (ATRESS) phenomenon (4), which may be caused by the selection and emergence of treatment-resistant cancer cells. Although it has been shown that therapy-treated metastases are strongly enriched for drug resistance mutations compared with untreated metastases, we believe another, possibly more important, explanation has been overlooked. Specifically, is it possible that the MTD of chemotherapy may cause tissue injury and activate cellular stress and senescence programs in cells in the nontumor environment outside the cancer? These extratumoral effects may remodel the landscape of distant nontumor environments and reshape cell–cell interactions in the metastatic niche and select for dormant, drug-resistant cancer cell states.

Cytotoxic chemotherapies have broad systemic effects on the body. Even at sublethal doses, they have the potential to affect specific cells and molecular processes beyond their recognized toxicities. These impacts are often underappreciated or perhaps minimized in contrast to the theoretical gains of eliminating residual cancer cells by prescribing higher-dose chemotherapy. We would argue that it may be impossible to eliminate every single cancer cell given the phenotypic plasticity, and pursuing this goal may have adverse effects on the behavior of cancer cells that disseminate early; select for drug-resistant dormant cell states; and promote subsequent disease progression at metastatic sites. Take as an example the skeleton, one of the most common sites of distant disease relapse. The impact of the chemotherapies used to treat primary disease on the cells of bone is poorly understood, and the effect this has on relapsing disease is even more limited.

At one level it has long been established that cytotoxic drugs may deplete bone marrow hemopoietic stem cells (HSC) and “empty out” the bone HSC niches. This has direct adverse effects on the innate and adaptive immune compartments and the ability of the immune system as a whole to target and eradicate cancer cells. However, chemotherapies also impact the process of bone turnover, often by stimulating osteoclastic bone resorption and inhibiting osteoblastic bone formation. These changes alter cells of the osteoblast lineage, including mesenchymal stromal cells (MSC), that make up the specific microenvironments, or niches, that support the long-term dormancy of disseminated cancer cells (5). In this regard, chemotherapies may act to facilitate the dissemination, colonization, and survival of cancer cells in specific niches in the skeleton and elsewhere.

Of greater importance is the impact of chemotherapies on disseminated cells that, at the time of diagnosis, have already seeded the skeleton and are residing in a dormant cell state. Osteoclastic resorption, by remodeling the dormant cancer cell niche in bone, can release cells from dormancy and cause disease relapse (6). Treatments such as melphalan, doxorubicin, androgen deprivation therapy, and aromatase inhibitors can all increase osteoclastic bone resorption. Targeting estrogen or androgen signaling pathways in breast and prostate cancer stimulate bone resorption and have been shown in preclinical studies to increase bone metastasis. This reflects the positive feedback amplification and vicious cycle of interdependence between cells in the growing tumor and the bone microenvironment.

On the plus side, there are widely used adjuvant treatments that inhibit osteoclastic bone resorption including the bisphosphonate, zoledronic acid, and the anti-RANKL mAb, denosumab. Bisphosphonates have been shown to maintain bone mineral density and reduce skeletal-related adverse events in women with breast cancer receiving adjuvant chemotherapy. In postmenopausal women, bisphosphonate use is associated with improved survival (7, 8). Adjuvant denosumab has also been shown to improve survival in postmenopausal women with hormone receptor–positive breast cancer (9). However, we still need to be mindful of how these bone-sparing agents are used. For example, stopping denosumab treatment for some patients is associated with accelerated bone loss and rebound fractures, possibly because of the rapid fusion of osteomorphs into actively resorbing osteoclasts (10). We suggest that such unrestrained osteoclastic bone resorption has the potential to remodel the endosteal bone surface and reawaken dormant cancer cells residing in that niche. Although the long-term outcomes from ABCSG-18 did not show an overall increase in rebound fracture risk for adjuvant denosumab compared with placebo, there were more vertebral fractures and more atypical femoral fractures. Therefore, it is imperative that the impact of adjuvant systemic therapies and their consequences on the EXTRA-TIME to be factored into the design of treatment regimes.

The impact of chemotherapy on the immune system, particularly in terms of anticancer immunity, is a neglected area that is only now beginning to receive the attention it deserves. The origin of this neglect may date back to 1976 with the adoption by the Developmental Therapeutics Program at the NCI of human subcutaneous xenografts into immunodeficient mice as a preclinical pipeline for in vivo anticancer drug testing. Although this arguably has been a highly successful drug development program, it has led to the testing of new drugs without consideration of how the cancer and drug candidates may interact with the immune system. More recently, the focus on the local TIME has drawn attention further away from cancer as a global failure of immune control. If we are to eliminate metastatic relapse and cure cancer, we need to consider cancer within the context of the entire immune system and the impact of all cancer therapy not just on the TIME but also the EXTRA-TIME outside the tumor in primary and secondary lymphoid tissues (Fig. 1).

The impact of chemotherapy on the immune system is now increasingly relevant because cancer immunotherapies, including immune checkpoint blockade (ICB), chimeric antigen receptor T cells, viral therapy, and anticancer vaccines, depend on the generation and propagation of effective anticancer immune responses. There is thus the challenge, on one hand, to minimize the immunosuppressive effects of chemotherapy, and the opportunity, on the other, to optimize the potential synergies between chemotherapy and immunotherapy to achieve durable cancer control (11). Patients receiving cancer chemotherapy are immunocompromised, make poor antibody responses to vaccines, and are at increased risk of life-threatening infections. Febrile neutropenia, sepsis, parasitic and other opportunistic infections, pneumonia, and influenza are common life-threatening complications of secondary immunosuppression in patients with cancer. However, chemotherapy may also cause immunogenic cell death and DNA damage and create neoantigens that have the potential to trigger anticancer immune responses. These potential benefits have yet to be harnessed to achieve responses in poorly immunogenic tumors and may overcome the poor response to ICB in breast cancer. Balancing these complex outcomes requires a deep understanding of all the individual components of the immune system and their precise coordination across multiple anatomic compartments in space and time to orchestrate local immune responses in the TIME and distal immune responses in the tumor-draining lymph node (tDLN). It is possible that this knowledge could drive sequencing of chemotherapies and immunotherapies that synchronize cytotoxicity with immunogenicity to achieve an optimal therapeutic outcome.

For example, alkylating agents such as cyclophosphamide and temozolomide given at high doses have profound antiproliferative effects that result in bone marrow suppression, neutropenia, and lymphopenia. In addition to these generic immunosuppressive effects, which are shared with some other chemotherapy agents, they also cause DNA damage and mutagenesis. Although this can promote cancer evolution, de novo carcinogenesis, and the development of secondary malignancies, such as leukemia and bladder cancers, it can also generate neoantigens that can be recognized as nonself and activate cancer-specific T and B cells to promote long-lasting anticancer immunity. Intriguingly, alkylating agents, in low doses, can deplete immunosuppressive regulatory T cells (Treg) that restrain humoral and cellular immune responses to promote anticancer vaccine responses in preclinical models. However, a single priming dose of cyclophosphamide did not significantly deplete Tregs or improve response to pembrolizumab in triple-negative breast cancer (12), suggesting that we need further refinement of these combination approaches. Indeed, a comparison of immune induction strategies showed that cisplatin and doxorubicin was superior to cyclophosphamide (13). DNA damage by alkylating agents induces a regulated form of necrosis that is inflammatory and potentially immunogenic. In contrast, other cancer chemotherapy drugs induce apoptotic cell death programs that are nonimmunogenic. Furthermore, tumor lysis may lead to the massive release of tumor antigens that can overload the immune system, lead to T-cell exhaustion, and induce high-dose antigen-specific T-cell tolerance. The importance of these global effects of cancer chemotherapy outside the tumor, which represent missed opportunities for mobilizing anticancer immunity, are also increasingly recognized for other cancer treatment modalities, including targeted therapies, radiotherapy, and surgery. There are, accordingly, numerous opportunities to thoughtfully combine these modalities.

The tDLN is regarded as an important stepping stone to hematogenous dissemination, and the identification of lymph node metastases is a critical component of every clinical cancer staging system. This has meant that lymph node dissection is often performed routinely as part of the surgical management. Could this perhaps be a misstep? The tDLN is also a critical staging ground for the priming and cross-priming of anticancer B and T cells. Interestingly, remodeling of the tDLN and histopathologic features such as the presence of reactive B-cell follicles, follicular hyperplasia, and germinal centers are often associated with good prognosis. In addition to metastatic tumor cells, the afferent lymphatics also deliver tumor antigens to the subcapsular sinus of the lymph node either as soluble antigen or transported by migratory dendritic cells and extracellular vesicles. There, they first encounter CD169⁺ macrophages that can present antigen to B cells and cross-present antigen to CD8⁺ T cells to generate antitumor immunity. Importantly, the priming of cytotoxic CD8⁺ T cells occurs in two stages, with an initial activation phase by dendritic cells in the tDLNs followed by acquisition of an effector differentiation program in the tumor (14). The tDLNs also serves as a reservoir of TCF1⁺ stem-like memory CD8⁺ T cells that replenish terminally differentiated effector T cells in the TIME (15). Preservation of the tDLNs is therefore likely critical to the success of ICB and other cancer immunotherapy approaches. Could this contribute to the survival benefit when resection of the tDLNs is delayed and ICB are given as neoadjuvant therapy? Evidence from breast-conserving therapy in the preimmunotherapy era, such as the ACOSOG Z0011 and SINODAR-ONE clinical trials, has shown that dissection of the axillary dissection does not confer any survival benefits in women with invasive breast cancer and sentinel node metastasis. Given complications such as lymphedema and the associated morbidity, should we abandon axillary lymph node dissection in breast and other cancers in light of this evidence?

In this commentary, we have introduced the concept of EXTRA-TIME to redirect attention to an important but overlooked aspect of cancer care. Cancer cell drug discovery has historically focused on targeting cell-intrinsic vulnerabilities that have prioritized measuring local tumor responses in immunodeficient in vivo models that have not only ignored the immune system but also ignored potential metastatic sites in the rest of the body. This has led to the clinical paradigm of finding and prescribing the MTD for cytotoxic chemotherapy, often to the detriment of the patient. There is evidence that a more nuanced, metronomic approach that uses less toxic, lower doses of chemotherapy may be noninferior and offers more potential for synergies with immunotherapy. Arguably, we have missed opportunities to exploit chemotherapy-induced immunogenic cell death and neoantigen generation to elicit potent long-lasting antitumor immune responses.

Understanding how EXTRA-TIME controls cancer progression, especially in the metastatic setting, will set the scene for the step-change needed to harness the immune system towards truly curative therapies. In parallel, understanding how current therapies that debulk tumors impact the nontumor immune environment will lead to a clearer understanding of how to sequence and schedule chemotherapy doses to maximize antitumor effects while minimizing long-term irreversible changes on normal cells in the body and how that, in turn, impacts disseminated cancer cells.

What are the lessons and way forward? We believe a more holistic approach to drug discovery, using immune competent in vivo models and evaluating the impacts of candidate drugs not just on the tumor volume but on the entire organism, including common metastatic sites, is mandated in the modern era. In the clinical trials space, there are also numerous opportunities to reimagine how we dose, schedule, and administer chemotherapy, targeted therapies, cellular therapies, radiotherapy, and theranostics in combination with immunotherapies. This may include trials involving using metronomic chemotherapy at below the MTD to induce a favorable TIME and EXTRA-TIME and thereby sensitize the tumor to immunotherapy. It may involve neoadjuvant ICB and using conservative surgery, leaving the tDLN intact, and low-dose chemotherapy to debulk, rather than completely eliminate, the tumor and provide a source of neoantigens to induce anticancer effector and memory T and B cells. It may also involve using genomics and machine learning approaches to personalize the type and dose of systemic therapy that will synergize best with the state of an individual's immune system. Some of these ideas that need to be tested, such as abandoning routine resection of the tDLNs, may seem heretical, but it is time we explore the EXTRA-TIME and develop the evidence base to change practice to combine and sequence all these modalities in a way that optimizes tumor debulking with the mobilization and execution of anticancer immunity.

P.I. Croucher reports grants from Relation Therapeutics and Angitia Biopharmaceuticals. C.L. Chaffer reports nonfinancial support and other support from Kembi Therapeutics outside the submitted work. No disclosures were reported by the other authors.

We thank Lisa Carey, Yuki Honda Keith, and Sarah Jacques for insightful comments and discussion. This work has been supported by the Australian National Health and Medical Research Council (NHMRC) Investigator grants APPID1155678 (to T.G. Phan) and APPID2009010 (to P.I. Croucher) and Ideas Grant GNT2019437 (to C.L. Chaffer); Mrs. Janice Gibson and the Ernest Heine Family Foundation; the Kinghorn Foundation; the Nelune Foundation; the UNSW Cellular Genomics Futures Institute; Cancer Council NSW; Leukemia Research Foundation; Australian Cancer Research Foundation (ACRF); and the Tour de Cure. K.N. Weilbaecher is supported by NCI-R01 R01 CA216840.