Natural killer cells

Natural killer cells are the primary anti-tumour cell type of the innate immune system. To differentiate between healthy and malignant cells, natural killer cells use various cell surface receptors to interrogate host cells and assess them for signs of viral infection or dysregulation common to cancer cells. If aberrant production of proteins is occurring in a cell, abnormal cell-surface protein expression profiles are detected by natural killer cells, and this initiates a series of processes that culminate in the destruction of the identified tumour cell (Moretta Reference Moretta, Bottino and Mingari2002). Additionally, natural killer cells have the ability to target and destroy tumour cells coated with antibodies, in the event that a humoral immune response to a tumour antigen has occurred (Malmberg Reference Malmberg, Carlsten and Bjorklund2017).

For a natural killer cell to kill a cancer cell, two mechanisms are primarily employed. First, the detection of a dysregulated or antibody-coated cell triggers the release of several effector molecules, including perforin and granzymes (Voskoboinik Reference Voskoboinik, Whisstock and Trapani2015). Perforin creates small pores in the cellular membrane of the targeted tumour cell, and granzymes can then pass through the holes made by perforins (Voskoboinik Reference Voskoboinik, Whisstock and Trapani2015). From there, granzymes initiate apoptotic signalling pathways through caspase cleavage, which leads to the death of the tumour cell (Voskoboinik Reference Voskoboinik, Whisstock and Trapani2015). The second major mechanism that natural killer cells use to induce cell death is through tumour necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL) signalling (Prager Reference Prager and Watzl2019). TRAIL is a membrane-bound protein on the surface of natural killer cells that binds the death receptors DR4 and DR5 (Prager Reference Prager and Watzl2019). If present on a tumour cell, death-receptor signalling then leads to apoptosis through caspase cleavage in a mechanism somewhat similar to granzyme-induced apoptosis (Prager Reference Prager and Watzl2019).

Although acute sympathetic activation, such as that observed during exercise, may be beneficial for improving natural killer cell mobilisation and function (Schedlowski Reference Schedlowski, Jacobs and Stratmann1993; Idorn Reference Idorn and Hojman2016; Pedersen Reference Pedersen, Idorn and Olofsson2016), early studies assessing the impact of repeated activation of β₂-adrenergic receptors (β₂-AR) on natural killer cells found it to be immunosuppressive and to lead to progression of a natural killer cell-sensitive tumour model (Shakhar Reference Shakhar and Ben-Eliyahu1998). As more sophisticated methods of investigation have developed, an increasing body of evidence suggests that exposure to physiologically relevant amounts of chronic stress signalling leads to suppression of natural killer cell cytotoxicity. These studies found that causing stress in mice or rats by various methods resulted in an increase in both sympathetic stress and glucocorticoid stress, which resulted in a decrease in killing of B16-F10 melanoma cells (De Lorenzo Reference De Lorenzo, de Oliveira Marchioro and Greco2015) and MADB106 tumour cell clearance (Rosenne Reference Rosenne, Sorski and Shaashua2014) respectively. Interestingly, on further interrogation, it was discovered that glucocorticoid signalling drove an increase in β₂-AR expression, thus increasing the sensitivity of natural killer cells to sympathetic stress (Rosenne Reference Rosenne, Sorski and Shaashua2014). Importantly, pharmacological inhibition of adrenergic signalling reversed the effect of not only sympathetic suppression, but also glucocorticoid suppression, suggesting that adrenergic signalling contributed most significantly to this phenomenon (Rosenne Reference Rosenne, Sorski and Shaashua2014).

Work from Thornton et al (Reference Thornton, Andersen and Crespin2007) evaluated natural killer cell cytotoxicity over time in women diagnosed with breast cancer, and they found that women who reported a more rapid reduction in stress in the weeks after diagnosis showed a greater improvement in natural killer cell cytotoxicity compared with women who reported a slow decline in stress levels (Thornton Reference Thornton, Andersen and Crespin2007). When cytotoxic activity was assessed in natural killer cells from the tumour microenvironment and blood of women being treated for ovarian cancer, it was also found that high social support correlated with increased cytotoxicity, and self-reported distress was associated with decreased cytotoxicity (Lutgendorf Reference Lutgendorf, Sood and Anderson2005).

Dendritic cells

Dendritic cells are another innate immune cell type that play an essential role in the generation of an anti-tumour immune response and are negatively affected by stress. As a diverse group of professional antigen-presenting cells, dendritic cells are responsible for sampling antigens present in the tumour microenvironment and initiating a robust adaptive immune response (Wculek Reference Wculek, Cueto and Mujal2020). During antigen uptake, dendritic cells assess the cytokine milieu and detect damage-associated molecular patterns (DAMPs) present in the area. This information is then used to provide contextual information to other cells about the environment from which the antigens came and promote escalation of the immune response or antigen tolerance (Wculek Reference Wculek, Cueto and Mujal2020).

When considering the implications of stress on dendritic cell function, it is important to note that adrenergic stress has been shown to limit cytokine production from dendritic cells after exposure to lipopolysaccharide (Goyarts Reference Goyarts, Matsui and Mammone2008), impair migration (Maestroni Reference Maestroni2000) and chemotaxis (Maestroni Reference Maestroni and Mazzola2003) and drive dendritic cells to preferentially promote differentiation of CD4⁺ T cells into type 17 helper T cells (Th17) rather than the type 1 (Th1) cells (Manni Reference Manni, Granstein and Maestroni2011) that are important for anti-tumour immunity. Although these observations did not take place during cancer studies, implications of these results suggest that β-adrenergic signalling may lead to the impairment of T-cell dependent anti-tumour immune responses initiated by dendritic cells (Knochelmann Reference Knochelmann, Dwyer and Bailey2018).

During the development of dendritic cells, glucocorticoid exposure has been shown to impair maturation, leading to poor antigen presentation and decreased cytokine production (Piemonti Reference Piemonti, Monti and Allavena1999). Rozkova et al (Reference Rozkova, Horvath and Bartunkova2006) assessed the impact of glucocorticoid exposure during both dendritic cell differentiation and maturation, and discovered further evidence suggesting an immunosuppressive role of glucocorticoid stress, although the magnitude of these findings appears to be culture-condition dependent (Vieira Reference Vieira, Kaliński and Wierenga1998; Canning Reference Canning, Grotenhuis and de Wit2000). Additionally, their ability to activate Th1 cells was shown to be limited, leading to increased development of immunosuppressive regulatory T cells (Tregs) (Matyszak Reference Matyszak, Citterio and Rescigno2000).

Subsequent investigations using a murine breast cancer model determined that, when adrenergic stress is induced by chronic cold stress, dendritic cell populations become more immature and less capable of stimulating CD8⁺ T-cell proliferation (Kokolus Reference Kokolus, Spangler and Povinelli2014). Additional work reported a similar decrease in maturation of antigen-specific dendritic cells when mice were exposed to social disruption stress during treatment with a cancer vaccine to a murine model of melanoma (Sommershof Reference Sommershof, Scheuermann and Koerner2017). When Botta & Maestroni used a tumour antigen-specific dendritic cell vaccine strategy to target a thymoma cell line expressing the ovalbumin antigen (EL-4 transfected with OVA), they found that an intradermal injection of the β²-AR agonist salbutamol impaired the efficacy of the anti-tumour dendritic cell vaccine treatment (Botta Reference Botta and Maestroni2008). Interestingly, although blockade of this signalling improved vaccine efficacy when immature dendritic cells were used, if the same procedure was carried out with more mature dendritic cells, the vaccine was ineffective and tumour growth progressed similarly to that in control mice treated with phosphate-buffered saline (Botta Reference Botta and Maestroni2008).

T cells

If a dendritic cell is going to stimulate an anti-tumour immune response, it will need to travel out of the tumour microenvironment and into a nearby lymph node to present tumour antigens to other immune cells (Chen Reference Chen and Mellman2013). One of the key populations responsible for directing this immune response are the T cells, which can generally be broken down into CD8⁺ and CD4⁺ T cells. CD8⁺ T cells are primarily responsible for the direct killing of targeted cells, whereas CD4⁺ T cells participate in coordinating the immune response. In the lymph node, naive CD4⁺ T cells differentiate into various mature populations, primarily helper T cells (Th) or regulatory T cells (Tregs), guided by input from antigen-presenting dendritic cells (Chen Reference Chen and Mellman2013). Tregs are immunosuppressive, whereas helper T-cell populations typically fall into one of two categories: Th1 cells promote the cell-mediated immune response key to anti-tumour activity and Th2 cells promote antibody production and the humoral immune response. Other populations, such as Th9 or Th17 cells, can also develop, but their implications on the anti-tumour immune response are conflicting and less characterised (Yan Reference Yan and Richmond2020), especially in the context of chronic stress, and therefore will not be discussed here.

During Th1 responses, cytotoxic CD8⁺ T cells are activated and are known to be a major component of effective anti-tumour immunity (Disis Reference Disis2010). They exert their effects on tumour cells through mechanisms similar to those used by natural killer cells, but with several key distinctions. Rather than targeting general signs of dysregulation, they only attack cells bearing specific antigens. To do so, they primarily employ two general mechanisms to kill a tumour cell. First, these T cells utilise the same effector molecules, perforin and granzymes, used by natural killer cells to permeabilise the cell's membrane and induce apoptosis (Harty Reference Harty, Tvinnereim and White2000). Additionally, CD8⁺ T cells can present a membrane-bound protein called Fas ligand (FasL) to the targeted cell, which may be expressing the Fas death receptor, and signalling downstream of the Fas death receptor should then also drive apoptotic pathways within the tumour cell, resulting in its death (La Reference La, Tai and Lee2009).

Interestingly, when treatments with the β-AR agonist isoprenaline were administered daily to mimic chronic stress conditions, the capability of CD8⁺ T cells to directly kill tumour cells was impaired, along with their ability to produce interferon-gamma (IFN-γ) (Nissen Reference Nissen, Sloan and Mattarollo2018). This led to a decrease in the efficacy of several immune-based therapies targeting a mouse model of lymphoma (Eμ-Myc B-cell lymphoma), including anti-PD-1 and anti-4-1BB monoclonal antibodies and a tumour antigen vaccine (Nissen Reference Nissen, Sloan and Mattarollo2018). Similar findings were discovered when a social disruption stress model was used to show that chronic stress impaired the immune response to melanoma vaccine by reducing IFN-γ production and killing of antigen-specific CD8⁺ T cells (Sommershof Reference Sommershof, Scheuermann and Koerner2017). Using the B16 model of melanoma and the 4T1 model of mammary carcinoma, Bucsek et al showed that the reduction of β-adrenergic signalling by knocking out the β₂-AR, treating mice with the β-AR antagonist propranolol or relieving chronic thermal stress with warmer housing conditions resulted in an improvement in anti-PD-1 monoclonal antibody therapy, driven by enhanced CD8⁺ T cell function, as evidenced by increased production of IFN-γ and granzyme B (Bucsek Reference Bucsek, Qiao and MacDonald2017).

For a CD8⁺ T cell to obtain its full anti-tumour potential, dramatic metabolic changes must take place during activation to allow for the increased energy demands required for rapid expansion and production of effector molecules (Siska Reference Siska and Rathmell2015). Interestingly, when CD8+ T cells were exposed to a pan β-AR agonist (isoproterenol) during differentiation, the necessary metabolic changes, such as increased glycolysis and glucose uptake, were unable to occur (Qiao Reference Qiao, Bucsek and Winder2019). This suggests that impaired metabolic reprogramming of CD8+ T cells is another mechanism through which β-adrenergic signalling promotes immunosuppression (Qiao Reference Qiao, Bucsek and Winder2019).

In addition to SNS signalling, glucocorticoid stress has also been shown to have profound effects on T-cell function (Ashwell Reference Ashwell, Lu and Vacchio2000), as one of many mechanisms attributed to T cells’ capacity to drive immune suppression (Kadmiel Reference Kadmiel and Cidlowski2013). By simply exposing mature T cells to glucocorticoids, Brunetti et al showed that apoptosis could be induced in this population (Brunetti Reference Brunetti, Martelli and Colasante1995), in addition to effects seen in less differentiated subsets of T cells by other researchers. Such effects include a skewing of the Th1:Th2 ratio towards a Th2 response (Ramirez Reference Ramirez, Fowell and Puklavec1996; Franchimont Reference Franchimont, Galon and Gadina2000) and the downregulation of FasL (Yang Reference Yang, Mercep and Ware1995), which should lead to impaired anti-tumour immunity.

It is also important to appreciate that stress caused by models not specific to HPA or SNS signalling, such as the chronic restraint of mice, leads to an increase in tumour growth driven by a significant reduction in circulating CD4⁺ T cells, a decrease in the production of the key Th1 cytokines IFN-γ and TNF-α, and the impairment of direct lymphocyte cytotoxicity (Frick Reference Frick, Arcos and Rapanelli2009). Interesting work by Hou et al described another novel model in which repeated exposure to the sound of a scream was used to cause chronic stress. They found that mice exposed to the screaming sound had elevated levels of both corticosterone and noradrenaline, and these mice experienced increased colon cancer growth, increased Th2 cytokine production and decreased Th1 cytokine production (Hou Reference Hou, Zhang and Zhao2013).

Importantly, human clinical trials that aimed to decrease stress through psychosocial telephone counselling similarly found that when patients received counselling during treatment for cervical cancer, they not only had improved measures of quality of life, but also experienced a shift in Th1/Th2 cytokine production towards Th1 (Nelson Reference Nelson, Wenzel and Osann2008).

Regulatory T cells

To constrain and prevent excessive immune responses, regulatory T cells (Tregs) are present throughout the body (Romano Reference Romano, Fanelli and Albany2019). Although Tregs are essential for the prevention of pathological processes such as autoimmunity, their involvement in an anti-tumour immune response decreases cytotoxic T-cell activity, thus promoting tumour growth (Zou Reference Zou2006). Tregs are CD4⁺ T cells that develop either in the thymus or the periphery and limit cytotoxic T-cell abundance and effector functions through various processes (Romano Reference Romano, Fanelli and Albany2019).

The first of three primary mechanisms by which Tregs limit the immune response is the expression of cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) (Corthay Reference Corthay2009). CTLA-4 is a cell-surface protein that binds a protein (B7) expressed on antigen-presenting cells, thus limiting proper co-stimulation of naive T cells and preventing the maturation of naive T cells into mature cytotoxic T cells (Corthay Reference Corthay2009). Second, Tregs express the receptor CD25, which is capable of binding and sequestering the cytokine interleukin-2 (Chinen Reference Chinen, Kannan and Levine2016). Interleukin-2 is an important cytokine responsible for promoting cytotoxicity of T and natural killer cells and for the proper development and activation of many other immune cells, including CD8⁺ memory T cells, B cells and macrophages (Bachmann Reference Bachmann and Oxenius2007). Finally, Tregs produce several immunosuppressive cytokines, such as transforming growth factor beta (TGF-β) and interleukin-10, which are responsible for decreasing T-cell proliferation and inhibiting the production of pro-inflammatory cytokines respectively (Corthay Reference Corthay2009). All of the above functions impair anti-tumour immunity, so changes in the number of Tregs can have dramatic consequences on anti-tumour immunity.

When chronic subordinate-colony housing techniques were used to induce chronic stress in mice, transplanted syngeneic fibrosarcoma tumours grew at a greater rate and an increase in Tregs was discovered (Schmidt Reference Schmidt, Peterlik and Reber2016). In other studies evaluating the impact of both acoustic stress and chronic restraint stress on a pancreatic mouse model, it was found that not only did stress increase the number of Tregs in the tumour microenvironment, but there it also reduced the production of both Th1 and Th2 cytokines and increased the production of TGF-β (Partecke Reference Partecke, Speerforck and Kading2016). In a series of studies evaluating the impact of radical mastectomy surgery on neuroendocrine signalling in humans being treated for primary breast cancer, Zhou et al found that levels of adrenaline and noradrenaline increased after surgery, along with the number and immunosuppressive capacity of Tregs (Zhou Reference Zhou, Li and Li2016). Importantly, this increase in Treg number and function could be abrogated with propranolol treatment (Zhou Reference Zhou, Li and Li2016).

Myeloid-derived suppressor cells

Myeloid-derived suppressor cells (MDSCs) are a heterogeneous group of myeloid cells that become prevalent and functional during various pathological inflammatory processes, primarily in the context of various cancers. This collection of relatively immature myeloid cells is identifiable in mice and humans on the basis of cell surface markers, and they promote immune suppression through their interactions with other cells in the tumour microenvironment. MDSCs are capable of suppressing the immune system, primarily by acting on T cells, through an abundance of mechanisms, including arginine metabolism by arginase I and inducible nitric oxide synthase (iNOS), reactive oxygen species (ROS) production, prostaglandin E₂ production via cyclooxygenase 2 expression (Take Reference Take, Koizumi and Nagahisa2020), TGF-β and interleukin-10 release, and indoleamine 2,3-dioxygenase synthesis (Groth Reference Groth, Hu and Weber2019), and therefore if stress leads to an increase in MDSCs, this is suggestive of impaired anti-tumour immunity.

In addition to the increased Treg development that occurred when mice were exposed to chronic subordinate-colony housing, this form of chronic stress also increased the accumulation of MDSCs, further contributing to the immunosuppressive environment that resulted in increased fibrosarcoma growth (Schmidt Reference Schmidt, Peterlik and Reber2016). Studies of hepatocellular carcinoma revealed that chronic restraint stress resulted in a redistribution of myeloid cells from the spleen to the tumour, causing increased tumour growth that could be inhibited with propranolol treatment (Jiang Reference Jiang, Li and Wei2020). Mohammadpour et al showed that increased β₂-AR signalling in mice resulted in increased growth of 4T1 and AT-3 mammary carcinomas and an increase in the accumulation and immunosuppressive function of MDSCs (Mohammadpour Reference Mohammadpour, MacDonald and Qiao2019), and Mundy-Bosse et al found that patients reporting higher levels of stress had increased levels of MDSCs that correlated with increased salivary cortisol levels (Mundy-Bosse Reference Mundy-Bosse, Thornton and Yang2011).

Macrophages

Under normal physiological conditions, macrophages are responsible for phagocytosing pathogens and debris, remodelling microenvironments and maintaining tissue homeostasis (Ostuni Reference Ostuni, Kratochvill and Murray2015). However, when macrophages are recruited to the tumour microenvironment, factors such as hypoxia, low pH and cellular debris from dying tumour cells drive tumour-associated macrophages (TAMs) to aid in tumour growth, neovascularisation, metastasis and immunosuppression (Ostuni Reference Ostuni, Kratochvill and Murray2015). Many of the immunosuppressive capabilities of macrophages are similar to those of MDSCs, and this includes production of interleukin-10, TGF-β and prostaglandin E₂, breakdown of L-arginine by arginase 1 and iNOS, and expression of programmed death ligand 1 (PD-L1) (Ostuni Reference Ostuni, Kratochvill and Murray2015).

Qin et al showed that an increase in adrenergic tone from social isolation of mice resulted in increased M2 macrophage polarisation and increased growth of 4T1 tumours (Qin Reference Qin, Jin and Li2015). Other work studying macrophage recruitment found that elevated adrenergic stress resulted in an increase in monocyte chemotactic factor 1 (MCP1), which led to an increase in macrophage recruitment to the tumour and poor survival (Armaiz-Pena Reference Armaiz-Pena, Gonzalez-Villasana and Nagaraja2015). Additional work studying the impact of adrenergic stress on lung macrophage populations found that stress promoted a pro-tumour macrophage population in the lungs that increased rates of metastasis (Chen Reference Chen, Liu and Guo2018) and increased M2 polarisation of macrophages in the tumour microenvironment (Sloan Reference Sloan, Priceman and Cox2010).