A review of immune therapy in cancer and a question: can thermal therapy increase tumor response?

Abstract

Immune therapy is a successful cancer treatment coming into its own. This is because checkpoint molecules, adoptive specific lymphocyte transfer and chimeric antigen T-cell (CAR-T) therapy are able to induce more durable responses in an increasing number of malignancies compared to chemotherapy. In addition, immune therapies are able to treat bulky disease, whereas standard cytotoxic therapies cannot treat large tumour burdens. Checkpoint inhibitor monoclonal antibodies are becoming widely used in the clinic and although more complex, adoptive lymphocyte transfer and CAR-T therapies show promise. We are learning that there are nuances to predicting the successful use of the checkpoint inhibitors as well as to specific-antigen adoptive and CAR-T therapies. We are also newly aware of a here-to-fore unrealised natural force, the status of the microbiome. However, despite better understanding of mechanisms of action of the new immune therapies, the best responses to the new immune therapies remain 20–30%. Likely the best way to improve this somewhat low response rate for patients is to increase the patient’s own immune response. Thermal therapy is a way to do this. All forms of thermal therapy, from fever-range systemic thermal therapy, to high-temperature HIFU and even cryotherapy improve the immune response pre-clinically. It is time to test the immune therapies with thermal therapy in vivo to test for optimal timing of the combinations that will best enhance tumour response and then to begin to test the immune therapies with thermal therapy in the clinic as soon as possible.

Keywords:

• Thermal therapy
• antigen-specific adoptive T-cell therapy
• CAR-T therapy
• microbiome
• checkpoint inhibitors

Introduction

Cancer treatment is at last becoming successful. One reason is that new immune therapies are proving more effective than most standard oncological treatments. The durability of response to immune therapy is significantly longer than the durability of responses to most chemo- or targeted therapies. In addition, unlike cytotoxic therapies that fail in the face of heavy tumour burden, immune therapies, including checkpoint inhibitors, antigen-specific adoptive therapy and CAR-T therapies induce responses that can effectively treat both metastatic and bulky disease. Importantly, today’s cancer immunotherapies have toxicities that are generally better tolerated than the toxicities of traditional cytotoxic cancer therapies [1]. In February 2017, for the second consecutive year, the American Society of Clinical Oncology (ASCO) designated immunotherapy as the cancer advance of the year [2]. For many years TNM staging and cancer histopathological type and grade have been the criteria to predict tumour response to treatment and overall survival. In fact, malignant tumour prognosis is the result of a balance between the oppositional challenge of host immune and inflammatory responses vs. the aggressiveness of a malignancy [3]. Today there is an increasingly accepted concept that the “immune index”, e.g., the number of effector lymphocytes in the tumour and its microenvironment are more important than either tumour staging or histopathological diagnosis for 5-year survival prognosis [3,4].

History

The recognition that the immune system plays an important role in controlling cancer seems like a new concept, yet immune therapy to treat cancer was first tested more than 125 years ago. In 1891, an American surgeon, William Coley injected live Streptococcus pyogenes into patients’ tumours. He had read German publications that described unexpected tumour responses in patients with high fevers from streptococcal infections. Coley later developed a vaccine made of a mixture of killed bacteria. His bacterial vaccines became known as “Coley’s Toxins” [5]. We consider William Coley to be the father of immunotherapy because his vaccines achieved remarkable, complete and durable tumour responses in patients with inoperable sarcoma. Yet at the time, physicians ignored Coley’s work because they were fearful of injecting cancer patients with pathogenic bacteria in a time antedating the development of antibiotics. In addition, Coley’s colleagues were sceptical because there was no explanation for the tumour response to “Coley’s Toxins”. Consequently, during the first part of the 20th century, cancer treatment continued to rely only on surgery and the “new” radiation therapy [5–7]. Interestingly, in the early 1900 s Coley developed more purified vaccines that did not cause high fever. However, few tumour responses occurred with the more purified vaccines, suggesting that fever was important to tumour response.

Continuing a role for bacteria in tumour therapy, forty years ago preparations of bacille Calmette-Guérin (BCG) were first instilled directly into the bladders of patients with bladder cancer. The BCG therapy proved to be a successful treatment in many patients with non-invasive disease. BCG continues to be used to treat non-muscleinvading bladder cancer [8].

Regarding the interplay of bacteria with tumour response, we may have come full circle. In a very different arena involving bacteria, gut bacteria have recently been shown to play a highly important role in immunity and specifically in response to cancer therapy. Recent data shows that commensal gut bacteria (microbiota) play a key role in tumour response through a profound impact on host immunity. Gut microbiota influence both tumour growth as well as response to chemo-, epigenetic and immune therapies [9–14]. In bone marrow transplantation, they play an important role in graft vs. host disease [13,15].

Although, now taken for granted, no one recognised the role lymphocytes played in the immune response until the 1960 s, when it was first demonstrated that T-lymphocytes mediated allograft rejection in mice [16]. Also, cytokines were discovered in the 1960s. One can posit that the discovery of the role lymphocytes, as well as cytokines play in immune response led to modern immunotherapy. One of the most important cytokines was interleukin-2 (IL-2) [17]. Using IL-2, Rosenberg demonstrated for the first time since Coley that immune therapy could effectively treat patients with malignancies. His clinical studies using high doses of IL-2 induced some complete and durable responses when administered to patients with metastatic melanoma and renal cell carcinoma, whereas chemotherapy was ineffective in these malignancies [18]. The responses induced by IL-2 were profound and durable [19]. However, while IL-2 therapy appeared to be a major breakthrough for immune therapy, toxicities were severe and the overall response rate was relatively low [20,21]. Therefore, immune therapy was once again viewed as ineffective [21].

For more than 20 years, no one had an explanation for why responses to immune therapies such as IL-2 were so few and so sporadic. Today we have a better understanding of why tumour responses had been so limited to immune treatments. To understand immune therapy, it is important to realise that during the development of vertebrates and mammals, including humans, the immune system has evolved to react to and kill infectious organisms. The immune system has also developed a long-lived immune memory to prevent re-infection with previously encountered infectious organisms. The intrinsic immune system (NK cells, macrophages, etc.) and the adaptive immune systems (CD8 + T cells, CD4 + T cells, etc.) constantly search for infectious invaders, with viruses being particular targets [22].

The host immune cells also search for diseased, or cancerous cells [23]. However, unlike infectious invaders such as bacteria, fungi, or viruses, which are “foreign” cells that differ completely from normal host cells; malignant tumour cells are derived from normal cells that is mutated, yet retain some genetic features of the normal cells, e.g., “self” cells. Tumor cells are more difficult to recognise as “other” or “foreign”, although in fact, the immune system can recognise mutated cells as different from normal cells, e.g., “non-self”. Host effector lymphocytes always recognise mutated antigens (neoantigens). Yet, it is apparent that immune effector cells usually do not react to malignant cells of established tumours. They do not react to the tumour cells because malignant tumours use a variety of strategies to elude or minimise detection by host immune cells and to prevent effector lymphocyte activation [24].

Tumor evasion of host immune destruction

The tumour repertoire of ways to elude immune recognition include down regulation or editing to remove tumour-antigen surface expression by the malignant cells [25,26]. Malignant tumours also down regulate the major histocompatibility complex-1 (MHC-1) [27]. MHC-1 is a cell surface protein on nucleated cells. The main function of the MHC molecule on immune cells is to bind to pathogen-derived neoantigens. Once bound, the MHC molecules then display the neoantigens for recognition by CD8 + and CD4 + Tcells [28]. The MHC-1 molecule is critical for cytotoxic CD8+ and CD4 + T cells to recognise foreign antigens. Malignant cell down-regulation of MHC-1 antigens decreases APC tumour-antigen presentation and results in a reduction in T-cell immune infiltrates in the tumour [27].

Additionally, both epithelial and hematological malignancies secrete immunosuppressive cytokines such as TGFβ, IL-10, IL-3 and immunosuppressive molecules such as arginase in the tumour microenvironment [29,30]. Other immune functions adopted by cancers to block host immunity include recruitment of host-derived inhibitory immune cells, e.g., CD3 + regulatory Tcells, myeloid suppressor cells and plasmacytoid dendritic cells. Malignant tumours recruit these inhibitory immune cells and use them to impede host cytotoxic lymphocytes. Malignant tumours also down-regulate NK cells [31–34]. Malignancies frequently use one or more of the above immune evasion processes to escape host immune surveillance.

In addition to the role that malignant tumour cells play, tumour stroma also blocks the host immune cytotoxic cells. Epithelial to mesenchymal transition (EMT) occurs in tumour stroma after chronic inflammation. EMT provokes cancer immune-regulatory effects that lead to a depletion of the number of T- and B-lymphocytes by inhibiting T-lymphocyte proliferation as well as causing T-lymphocyte and NK apoptosis [35,36]. Further, when effector T-cells are chronically blocked, they become exhausted. Exhausted T-cells are no longer able to recognise malignant cell neoantigens and cannot kill cancer cells [37].

An overarching concept is that an effective immune response from the host is critical for durable responses to any therapy, including immune therapy, chemotherapy, targeted therapy or radiation therapy [3]. Therefore, the above described protective tumour mechanisms to avoid detection or block host cytotoxic lymphocytes, in part explain why response to immune therapy has been so low. Diminished or absent immune surveillance by the host allows unimpeded proliferation of the malignancy [38].

Checkpoint inhibitors

Of all mechanisms that tumours use to evade a host immune response, malignant tumours most commonly appropriate host checkpoints inhibitors [23,39] to impede host effector lymphocytes [40,41]. The host uses checkpoints inhibitors to protect normal cells from autoimmune injury. In the normal immune system, antigen presenting cells (APCs) e.g., dendritic cells and macrophages, obtain foreign antigens, migrate to lymph nodes and “prime” naïve T-cells using specific infectious antigens or tumour cell neoantigens. The dendritic cell APCs use the MHC-1 molecules and an immune stimulating molecule, such as CD28, to prime the naïve T-cells. The APC priming of the naïve T-cells eventuates in mature cytotoxic CD8 + T-cells that can recognise and react to specific foreign proteins or neoantigens. The mature effector T-cells then migrate to the foreign cell and kill the cell displaying that foreign protein or mutated antigen. This protects the host and is the purpose of the immune system. However, normal cells, referred to as “self” cells, sometimes need shielding from immune attack. Because of this need to protect normal cells, in addition to stimulatory, “attack” co-receptors, the immune system also developed inhibitor molecules to prevent effector T-cells from harming normal, “self” tissue cells, e.g., non-infected, non-diseased or non-mutated cells. The purpose of the inhibitory co-receptor molecules is to prevent effector lymphocyte-induced autoimmune damage to normal tissue. The inhibitory or “stop” molecules are known as “checkpoints inhibitors”.

Historically, the first checkpoint inhibitor discovered was the cytotoxic T-lymphocyte associated protein-4 (CTLA-4) receptor molecule. During the 1990 s, it was widely believed that CTLA-4 stimulated immune effector T-cells. However an investigator, Dr. James Allison, disagreed with the scientific dogma of the time [42]. He found that CTLA-4 is an inhibitory not a stimulatory receptor. The CTLA-4 molecule on the surface of APCs actually acted as an “off” switch [19]. CTLA-4 acts early in the immune pathway by binding to effector T-cell ligands to prevent APC cells from programming naïve T-cells to become sensitised to the normal tissue. Allison was the first to find that tumours co-opted the CTLA-4 checkpoint inhibitor molecule and used CTLA-4 for protection from immune destruction [43–45]. He discovered that almost all malignant tumours use normal host checkpoint inhibitors to paralyse effector lymphocytes. The appropriated checkpoint molecules prevent cytotoxic lymphocytes from traversing the malignant microenvironment, entering the tumour itself and killing the tumour cells. Finally, Allison’s investigations made it clear that malignant tumours are responsible for inhibiting the host immune system and causing the sporadic response to immune therapies.

Allison’s experiments showed that by blocking the tumour-generated CTLA-4 checkpoint molecule, the previously paralysed effector CD8+ & CD4 + T cells were free to recognise the tumour, penetrate the tumour microenvironment, enter the tumour, then “touch” and kill the malignant cells [46–49]. Allison pioneered the first clinical use of an anti-CTLA-4 antibody in patients with melanoma. However, initially in the clinical trial, tumour responses to the anti-CTLA-4 inhibitor antibody were disappointing because responses were assessed using the same response time criteria used to evaluate tumour response after cytotoxic therapy (chemo-, targeted- or radiation-therapy).

Response to conventional cytotoxic therapy is measured using the Response Evaluation Criteria in Solid Tumors (RECIST) [50]. After treatment with conventional agents, tumours typically decrease within weeks. RECIST criteria assess tumour response by 8–12 weeks after beginning cytotoxic treatment and evaluate the treatment a failure (progressive disease, PD) if a tumour increases in size or new tumours appear in that time period [50]. Tumor reduction after an initial cycle of treatment can be predictive of improved survival [51].

Unlike RECIST tumour response to cytotoxic therapies, after immune therapies tumour regression often takes considerably longer than after conventional cytotoxic therapies. Allison demonstrated responses to the CTLA-4 inhibitor, blocker ipilimumab often occurred considerably later than 8–12 weeks. In fact, in response to immune therapies, tumours can initially enlarge before diminishing in size (“pseudo-progression”), and new lesions can appear. Actually, immune therapies induce four different types of responses. (1) Decrease in size of the baseline tumour(s), with no new lesions. (2) An initial stable disease followed by a gradual decline in tumour burden. (3) A decrease in tumour size after an initial increase in tumour burden and (4) Response in one lesion accompanied by an initial appearance of new lesions. All four patterns of response are associated with favourable survival [44]. After immune therapy, conventional “complete” (CR) and “partial responses” (PR) as well as “stable disease” (SD) often occur after an increase in tumour burden that conventional RECIST criteria would have labelled PD. There are immune-related response criteria (irRC) specific to immune therapies that correspond to RECIST criteria for cytotoxic therapies. An immune-related complete response (irCR) is the disappearance of all lesions and no new lesions. An immune-related partial response (irPR) is an eventual 50% drop in tumour burden from baseline as defined by the irRC. Immune-related progressive disease (irPD) is a 25% increase in tumour burden from the lowest level recorded. All else is considered immune-related stable disease (irSD). This important difference in response to immune therapy compared to cytotoxic therapy is a consequence of the biology of the immune system. For many patients, the immune system may get underway at a later time, even months after an immune treatment begins, even if tumour burden is initially increasing. After early tumour growth, an immune therapy can induce a subsequent meaningful decrease in tumour burden. A 25% threshold for increase in tumour size allows for this delay in response to immune therapies [52,53]. Despite an initial cynicism of researchers and clinicians, Allison persisted using his CTLA-4 checkpoint blocker antibody, ipilimumab. He was eventually able to demonstrate complete, durable responses in metastatic melanoma. The FDA approved the treatment for metastatic melanoma. His pioneer work with CTLA-4 led to the current burgeoning of checkpoint inhibitor therapy [54].

A second checkpoint inhibitor pathway includes the programmed death-1 (PD-1) receptor, along with the programmed death ligand-1 (PD-L1) and programmed death ligand 2 (PD-L2) molecules. The PD-1 pathway molecules appear later in the APC-immune effector T-cell cycle pathway than the CDLA-4 antibody [55]. The PD-1 pathway molecules act when the neoantigen-specific immune effector cells find the tumour cell bearing the recognised neoantigen(s) [56]. For protection of normal host tissue, induction and maintenance of effector T-cell tolerance requires PD-1. Its ligand PD-L1 on non-lymphocytic cells limits effector T-cell responses and protects normal cells from immune-mediated cell injury. In the normal host, PD-1 receptor is found on the surface of T-lymphocytes, B-lymphocytes, natural killer T-lymphocytes and activated monocytes as well as dendritic cells, the “professional” antigen presenting cells (APCs) [57]. PD-1 binds to its PD-L1 and PD-L2 ligands, expressed on APCs [58,59]. In a normal host, when effector T-cells encounter host-produced PDL-1 checkpoint molecules, the effector CD8 + T-cells remain inactive. The result of this is that when inflammation occurs in normal host tissue, the PD-1/PD-1 inhibitory pathway promotes tolerance of the inflamed organ to avert normal tissue damage [58,59]. Because tumour-infiltrating cytotoxic T-lymphocytes from patients with cancer often express PD-1, tumour-derived PD-1 checkpoint molecules inhibitors can directly block the host cytotoxic effector T-lymphocytes, preventing recognition of the tumour cells [60]. Tumor stroma, as well as malignant tumour cells often co-opt the PD-1 pathway molecules [61]. Tumor produced PD-1 checkpoint molecules effectively protect the tumour from the host cytotoxic immune response [57]. Additionally, many human epithelial and hematological malignancies express PD-1 and PD ligand-1 (PD-L1). Tumor expression of PD-1 and PDL-1 is associated with a worse prognosis, unless anti-PD-1 or anti-PDL-1 inhibitors are used [62].

To succinctly describe the difference between the CTLA-4 and the PD-1 inhibitors, CTLA-4 is expressed on the surface of immune system T-cells and acts as a kind of brake on the T-cells, preventing them from being activated to attack cancer cells. By blocking CTLA-4, the T-cells remain activated. Whereas PD-1 prevents activated T-cells from seeing and killing cancer cells that have the PD-ligand protein on their surface. When PD-1 binds to its ligand, PD-L1, on either normal or cancer cells, it prevents T-cells from attacking other cells in the body. PD-1 and PDL-1 inhibitors allow the cytotoxic lymphocytes to attack the cancer cells that express the PDL-1 ligand.

The FDA has currently approved five anti-checkpoint inhibitor drugs to treat patients diagnosed with seven different malignant diagnoses, including melanoma and non-small cell lung cancer (NSCLC), both squamous cell lung carcinoma and adenocarcinoma. The FDA has also approved anti-checkpoint inhibitors to treat renal cell carcinoma, head and neck cancer, Hodgkin’s disease, urethral cancer, Merkel cell cancer, gastric and liver cancer. The FDA approved ipilimumab (CTLA-4), nivolumab (PD-1), pembrolizumab (PD-1), atezolizumab (PDL-1) and avelumab (PDL-1), each for specific advanced malignancies.

As described above, the first checkpoint inhibitor blocker to be clinically tested was CTLA-4 (ipilimumab). In 2011 ipilimumab was first approved by the FDA to treat stage IV, metastatic melanoma and in 2015, the FDA broadened the use of ipilimumab to treat stage III melanoma. The second checkpoint inhibitor blocker that was FDA approved was an anti-PD-1 inhibitor, nivolumab. This drug was FDA approved in 2014 to treat advanced melanoma and later that year to treat persistent or recurrent NSCLC after progression chemo- or targeted therapy. The FDA approved nivolumab to treat both recurrent squamous cell cancer and adenocarcinoma. That same year the FDA also approved a second anti-PD-1 inhibitor, pembrolizumab, to treat previously treated adenocarcinoma and squamous cell carcinoma of the lung. Later, in October 2016, the FDA approved pembrolizumab for frontline treatment of patients with metastatic NSCLC whose tumours have more than 50% PD-L1 expression. Also in October 2016, the FDA approved the first PDL-1 inhibitor blocker, atezolizumab to treat patients with metastatic lung adenocarcinoma whose tumours have progressed after a platinum-containing regimen or after an anti-EGFR or ALK targeted therapy. In December 2016, the FDA approved a combination checkpoint inhibitor therapy using nivolumab combined with ipilimumab to treat advanced NSCLC. To treat NSCLC at this time, there are four anti-checkpoint inhibitors available, a CTLA-4 checkpoint inhibitor blocker, two PD-1 inhibitors blockers and two PDL-1 anti-checkpoint inhibitors [63]. In November 2015, the FDA approved nivolumab to treat advanced renal cell carcinoma. In 2016, the FDA approved nivolumab to treat relapsed classical Hodgkin’s disease and in March 2017, the it also approved pembrolizumab to treat relapsed Hodgkin’s disease. Also in 2016, the FDA approved both nivolumab and pembrolizumab to treat recurrent head and neck cancer [64]. Later in 2016, the PD-L1 anti-checkpoint inhibitor, atezolizumab was FDA approved to treat advanced urethral carcinoma

In March 2017, the FDA granted an accelerated approval to another PD-L1-inhibitor, avelumab, to treat patients with metastatic Merkel cell carcinoma, including previously untreated patients.

In September 2017 the FDA approved the use of pembrolizumab for patients with gastric or gastro-esophageal junction cancers that express PD-L1, if their cancers have progressed after prior chemotherapy. Also in September 2017 the FDA did an accelerated approval for nivolumab to treat hepatocellular carcinoma, a liver cancer caused by the hepatitis B and C viruses and cirrhosis. If the disease has progressed after sorafanib (nexavar) therapy. PD-L1 expression is not required. Again in September 2017, the FDA approved pembrolizumab in combination with chemotherapy (carboplatin + pemetrexate) for non-squamous, non-small cell lung cancer with or without PD-L1 expression.

Although not yet FDA approved, there are promising results reported from ongoing Phase III trials using nivolumab and atezolizumab with chemotherapy to treat advanced triple-negative breast cancer [65,66]. Patients with asymptomatic melanoma brain metastases treated with a combination of ipilimumab and nivolumab experienced a 53% overall response rate, and at nine months the median duration of response was not reached [67]. There are also promising results from early stage clinical trials using nivolumab in non-Hodgkin’s Lymphoma [68], small-cell lung cancer [69] and ovarian cancer [70]. The number of trials of anti-checkpoint inhibitors blockers for other malignant diagnoses are burgeoning.

Recently, a previously unrecognised checkpoint protein, CD38, has been reported. CD38 works by inhibiting CD8 + T cell function. Targeting this protein overcomes resistance to use of anti PD-1, PD-L1 inhibitor agents in lung cancer animal models. Preclinical work suggests a combination of anti PD-L1 with CD38 inhibition improves tumour response in animal models of lung cancer [71].

The checkpoint inhibitor blockers are unique among anticancer agents because unlike chemotherapy, radiation therapy or targeted agents, they do not specifically target malignant cells. Instead, the anti-checkpoints inhibitors target the tumour-derived molecules that tumours use to impair host immune T-cell function. To reiterate, the goal of checkpoint blockade therapy is not to activate the host T-cells but rather to remove the inhibitory constraints that prevent the host immune cells from reaching and killing the tumour cells [46,55]. However, as single agents, the anti-CTLA-4 and anti-PD-1 checkpoint inhibitors typically only induce response in 20–30% of patients [72]. It is very important for a patient with a malignancy that their physician find a way to predict whether that individual patient will respond to an anti-checkpoint inhibitor. In fact, there are robust response correlates to the use of anti-checkpoint inhibitors.

Factors that influence response to checkpoint inhibitors

The expression of PDL-1 on a tumour correlates with better response to PD-1 and PDL-1 inhibitors [73]. Topalian et al. demonstrated increased objective tumour responses with anti PD-1 inhibitors when tumour cells expressed PD-L1 [74]. Other trials have also demonstrated response correlation with PDL-1 tumour expression. None-the-less, despite that PDL-1 expression has become a criterion for inclusion of patients in studies of anti-PD-1 therapy [64], PDL-1 expression by the tumour does not absolutely correlate with response to drugs that block the PD-1 pathway. Some tumours do not express PDL-1, yet respond to PD-1 pathway inhibitor blocker therapies. Also, often malignancies that do express PDL-1 do not respond to the PD-1 pathway inhibitor agents [75].

Importantly, before being able to kill a malignant cell, effector lymphocytes must first be able to recognise it as abnormal. Immune cytotoxic T-cells can only identify malignant cells by recognising mutated peptides expressed on tumour cell surfaces [76]. A malignant cell can be recognised because it expresses mutated peptides bound to MHC-1 molecules on its surface. These mutated peptides are known as neoantigens and are not expressed on normal cells. Rizvi et al. demonstrated that response to anti-PD-1 inhibitors correlates with the numbers of neoantigens expressed by the malignancy [46]. During a clinical trial for patients with NSCLC using the anti-PD-1 inhibitor blocker, pembrolizumab, Rizvi et al. made what seemed a surprising observation at the time. Lung cancers in patients who smoked actually had a better response to pembrolizumab, compared to non-smokers. The investigators found that the reason lung cancer in smokers responded better to the anti-PD-1 inhibitor was that lung cancer in patients who actively smoked or had smoked, contain a significantly greater number of neoantigens compared to the lung cancers in non-smokers. The investigators were the first to demonstrate that response of NSCLC to pembrolizumab correlated with the number of neoantigens expressed on the tumour [77]. Anti-PD-1 inhibitor treatment of head and neck cancer also shows the same correlation [78,79].

The importance of neoantigen density for immune identification of tumour cells is also seen in cancers with microsatellite instability (MSI). Epithelial malignancies with MSI (MSI-high) harbour an elevated number of somatic mutations (neoantigens). MSI-high cancers have a greater response to anti-PD-1 therapies compared to cancers without MSI. The MSI-high cancers were associated with prolonged progression-free survival [80]. Microsatellite instability results from DNA mismatch-repair deficiency. Le et al. performed a study of the effect of pembrolizumab on patients with colorectal cancer and other malignancies comparing mismatch-repair deficient cancers to mismatch repair-proficient cancers [80]. In patients with mismatch-repair deficient cancers, the objective response rate was 40% (4/10) and progression-free survival was 78% (7/9), compared to 0% response rate (0/18 patients) in patients with mismatch repair-proficient cancers, respectively [80,81]. Whole-exome sequencing showed a mean of 1782 somatic mutations per tumour in mismatch-repair deficient tumours, as compared with 73 in mismatch-repair proficient tumours. This data again demonstrates that those cancers with high numbers of somatic mutations were associated with response to an anti-PD-1 inhibitor and had a prolonged progression-free survival [81]. In fact, the FDA has approved treatment with PD-1 inhibitors for any tumor that contains MSI hi.

Approximately 20% of all cancers worldwide are associated with viruses [82]. Virus-associated tumours express peptides that are completely different from the peptides expressed by normal cells. Because the virus antigens are foreign to the immune system, the immune system recognises the viruses [83,84]. Therefore, viral-associated malignancies may be more likely to respond to anti-checkpoint inhibitor therapy, as well as other immune therapies [85,86]. There is an emerging research area of treating virus-associated cancers with anti-checkpoint inhibitors and other immune therapies [87–90].

Immune index

Compared to density of neoantigens, an even more powerful predictor of response to anti-checkpoint inhibitors are the numbers of effector T-lymphocytes in the tumour and tumour environment. The number of immune effector T-cells appears be more important than a malignant tumour’s expression of a high number of tumour neoantigens [56], (or PDL-1 expression) [91]. The number of effector T-cells in the tumour and its microenvironment is a major component of the “immune index” of a malignant tumour [4,92]. Data suggests that the immune index may be the most important predictor of tumour response to immunotherapies but also predicts response to chemotherapy, targeted therapy and radiation therapy. Determining the numbers of cytotoxic CD8+, CD4+ and NK effector cells in the tumour or the tumour microenvironment unambiguously predicts whether tumours will respond to a checkpoint inhibitor. CD8 + T cell density at the invasive tumour edge correlates with response to anti-PD-1 checkpoint inhibitors [3,92,93]. The reason that PDL-1 expression at the invasive tumour edge correlates with a poor tumour prognosis to chemotherapy or targeted therapies is because PDL-1 has been co-opted by the malignant tumour to prevent host immune effector cells from entering the tumour or its stroma [94]. The immune system is important, indeed basic, even for tumour response to cytotoxic therapies.

Additionally, once host T-cells are in the tumour microenvironment, their ability to kill tumour relies on their competence to overcome a number of tumour-derived barriers. For recognition by the immune T-cells, the tumour must express MHC-I [27]. The effector-lymphocytes not only must recognise the malignant cell, but the effector T-cells must actually “touch” the tumour cells to kill them [46,56,95,96]. Enzymes must be available to cleave endothelial and other intracellular walls. Prakesh et al. have shown that granzyme B, released by leucocytes, cleaves multiple intracellular substrates necessary for a CD8 + Tcell to lyse target tumour cells. The enzyme promotes cytotoxic T-cell diapedesis into the tumour [97]. As previously described, there are other tumour-created barriers include tumour-co-opted regulatory T-cells, myeloid-derived suppressor cells and tumour-initiated inhibitory cytokines. These entities all impede antitumor immune responses. Additionally, if cytotoxic T-cells are continually blocked by chronic infections or malignancies, they become exhausted and are then ineffective in killing either infectious agents, or tumour cells [37].

Immune-deficient patients, are those patients who have received multiple chemotherapy treatments or are taking long-term corticosteroids, or organ-transplant patients or those that are genetically immune-deficient, have a paucity of effector T-cells in their tumour. This lack of effector immune cells in the malignancy may explain why checkpoint inhibitors and other cytotoxic anti-cancer regimens can be ineffective in heavily pre-treated patients. If effector T-cell numbers are scant in patients with cancer, what can be done to increase the number of effector lymphocytes in the tumour and its microenvironment is a topic of recent research.

Antigen-specific adoptive T-cell transfer

A robust approach is Rosenberg et al.’s use of antigen-specific adoptive transfer therapy of cytotoxic T-cell lymphocytes [98,99]. Despite widespread scepticism for immune therapy that began in the 1970 s and persisted throughout many frustrating years [19,100], Rosenberg persisted in his immune therapy research. Today his work has evolved to become a successful adoptive transfer of ingeniously modified, antigen-specific tumour-infiltrating lymphocytes (TILs). Adoptive transfer of antigen specific effector T-lymphocytes significantly increases the number of neoantigen-specific T-cells which are able to penetrate the tumour and tumour microenvironment to kill tumour cells [101]. Antigen-specific adoptive transfer is a therapy that can effectively treat not only melanoma, but also poorly responsive epithelial malignancies [102,103]. Adoptive T-cell therapy requires isolation of tumour-reactive T-cells or APCs exposed to a specific patient’s tumour antigens, followed by ex vivo expansion of the cells until a large enough number of T-cells are available for administration back into the patient. A conditioning chemotherapy is given prior to the adoptive T-cell administration. The need for the conditioning chemotherapy is because depletion of the host lymphocytes is a critical step for the transferred adoptive effector lymphocytes to persist and thus be effective. The mild conditioning regimen substantially increases infused cell persistence and the incidence as well as the duration of clinical responses [104]. Certainly antigen-specific adoptive T-cell transfer is a complex, expensive therapy; yet, it can elicit complete and durable responses in epithelial malignancies that have proven resistant to other standard and immune therapies. Notably, the accumulated bulk of a cancer does not appear to affect the efficacy of adoptive transfer immune therapy [30]. Despite its complexity, antigen-specific adoptive therapy has great promise.

CAR-T adoptive cell transfer

Adoptive transfer using chimeric antigen T-cell (CAR-T) therapy using the B-lymphocyte antigen, CD-19 is another highly promising immune therapy [105]. It has shown good responses in both paediatric and adult patients with acute lymphocytic leukaemia (ALL) and other B-cell malignancies. The FDA granted CAR-T breakthrough therapy approval for use in patients with ALL. In April 2017, the FDA also approved CAR-T (CD-19) treatment as a breakthrough therapy to treat patients with resistant diffuse large B-cell lymphoma. CAR-Ts consist of an antigen-binding domain and a T-cell activating domain, derived from a T-cell receptor, with a co-stimulation domain. The genetic material (in this instance, CD-19) is transferred into the T-cells by electroporation or viral transduction using a lentivirus or a gamma retrovirus. Unlike Rosenberg’s antigen-specific adoptive T-cell transfer therapy, which uses personalised antigen-specific T-cells for each individual patient, which is more complex, and more time-consuming to prepare; CAR-Ts are prepared by genetic antigen retargeting of bulk T-cells (currently using CD19) [105]. This allows the preparation of a large number of tumour-reactive T-cells in large-scale quantities in a relatively short time-period. The same preparation of CAR-T-cells can be administered to multiple patients with CD19 + B cell neoplasms. Like with antigen-specific adoptive therapy, a conditioning chemotherapy regimen to deplete the host lymphocytes must be administered before CAR-T therapy, to see long-term response. CAR-T therapy can have severe toxicities, including cytokine release syndrome (CRS), neurologic toxicity and B-cell aplasia [106]. The FDA approved CD19 CAR-T therapy for paediatric patients with ALL in July 2017. Despite its success in B-cell tumours, unlike antigen-specific adoptive T-cell transfer, CAR-T therapy has not yet proven useful in epithelial malignancies. Yet, research continues with CARs in epithelial malignancies [107,108].

The influence of the microbiome on immune function

The microbiome is an entirely different entity that significantly influences response to immune therapy. A rapidly emerging concept in cancer biology implicates the gut microbiota as a potent environmental factor that modulates the carcinogenic process [99,110]. Intestinal microbiota influence both cancer development and therapeutic response [9]. The microbiome affects immune therapy as well as cytotoxic chemo- and radiation therapy [9–11].

Microbial communities are commensal microbes. Microbiota is the name of these microbial communities. Microbiome refers to the microbial organisms and their genomes. The significant effects of the microbiota on carcinogenesis and cancer therapy illustrate the importance of the multifaceted relationship between the microbiome and the host [111]. The vast majority of microorganisms reside within the intestine and influence not only the local gut function but also exert long-distant effects on host well-being as well as disease [112]. Microbiota communities also reside in the skin and lung [113]. Preclinical models suggest that microbial dysbiosis has a causative impact on cancer development, including colorectal cancer and as well on other non-gut cancer diagnoses, including breast, lung, urogenital and liver cancer [114]. The explanation of why the gut microorganisms are so important to the host is that over the eons vertebrates have co-evolved with microorganisms, resulting in a symbioticand commensal relationship. Today we recognise that this symbiosis actually plays an important function in health and disease. The human body is a host to over 100 trillion symbiotic microbes residing in distinct locations of the body. We contain more commensal microbial organisms than the number of our own cells. Our symbiotic microbiota outnumber our own cells by 10 to 1 [113,115]. The gut microorganisms amount to as much as four pounds of biomass. Each individual person harbours nearly 1000 different types of microbes, primarily in the gut [116].

The gut commensal microbes sustain basic physiological processes: digestion, vitamin synthesis and host defence. Healthy individuals differ vastly in the types of microbes that colonise them. Every individual has a unique mix of microbial species [117]. Significantly large differences between individuals occur because of different exposures to microorganisms during development in utero as well as after birth and also depend on the environment, individual genetics, age, diet and exposure to antibiotics [113,117,118]. Importantly, variations in the microbiota between individuals influence tumour immunity [119]. The indigenous microflora stimulate the host immune system to respond to pathogen challenge [120]. Adaptive immunity is the part of our immune system that learns how to respond to microbes after first encountering them, enabling a more rapid defence against disease-causing organisms. Without contact with the microorganisms that colonise us from birth, our adaptive immunity would not exist. Germ-free rodents, because they lack this defence, have an underdeveloped immunity.

Specifically, the gut microbiota has a strong influence on the regulation and maturation of lymphoid tissues acts both locally and systemically to regulate the recruitment, differentiation and function of host innate and adaptive immune cells [121]. The composition of the commensal bacteria of the gut microbiota can significantly alter antitumor immunity and specifically affect response to anti-PDL-1 immunotherapy [9,10,122].

Antibiotics detrimentally affect microflora by causing a change in the microbiota that impairs immunity [115]. A single course of antibiotics can disrupt the normal makeup of microorganisms in the gut for as long as a year [123]. This becomes a significant problem for patients undergoing intense chemotherapy regimens because chemotherapy-induced neutropenia makes them susceptible to infection, for which they are treated with broad-spectrum antibiotics [13]. Additionally, long term or repeated antibiotic use, increases colon adenoma incidence. Adenomas carry an increased risk of colon cancer [124].

As described above, bacteria that live in the gut have dissimilar compositions in different people. However, does an individual’s microbiome composition correlate with responsiveness to immune therapy, is a question that needs to be analysed. Sivan et al. answered this in the affirmative during pre-clinical studies using a PD-L1 checkpoint inhibitor [19]. The researchers compared the growth of subcutaneous B16-SIY melanoma in genetically identical C57BL/6 mice derived from two different mouse facilities, abbreviated JAX and TAC. The genetically identical mice from these two facilities have different commensal gut microbes. The researchers found that the B16SIY melanoma growth rate was significantly different in JAX compared to TAC mice. Implanted melanomas grew more rapidly and were larger in TAC mice contrasted to JAX mice. The investigators also showed that PD-L1 therapy was significantly less efficacious in TAC mice than in JAX mice.

The investigators then performed faecal transplants, transferring JAX faecal material to the TAC mice. The transplant of JAX faecal material to the TAC mice, resulted in a significantly slower tumour growth, accompanied by increased tumour specific T-cell response and increased infiltration of antigen-specific T-cells into the tumours of the transplanted TAC mice. A combination treatment using both JAX faecal transfer with anti-PD-L1 therapy significantly increased tumour control in the TAC mice [9] Notably, the therapeutic effect of faecal transplant was abrogated in CD8 + lymphocyte-depleted mice. This data indicated that the mechanism of faecal microbiota effect on tumour therapy was not direct, but rather through host antitumor T-cell response.

To summarise, the therapeutic results of immune therapy can be correlated with the numbers and activity of pre-existing immune effector lymphocytes, the number of tumour mutations/viral gene codes and more recently recognised, the status of the host microbiome [125]. The above discussed impairments of immune response caused by several mechanisms of both tumour and host-related factors, suggest that we can increase immune response by manipulating these elements [125].

Yet, while there are available mechanisms we may be able to use to increase host immunity against malignancies in the future; today, despite impressive efficacy, only 20–30% of candidate malignancies respond either to single immune checkpoint inhibitors [73,126]. or to specific antigen T-cell transfer immune therapies [101]. To a patient with cancer, a 20–30% response rate is low. Enhancing the intrinsic immune response of the host is an appealing option to increase the number of responses.

The role of hyperthermia/thermal therapy in immune therapies

Thermal therapy is an under-investigated path of clinical research, that is highly likely to increase response to both anti-checkpoint inhibitors and other immunotherapies. The efficacy of immune therapies can be expected to increase when they are combined with thermal-therapy (“hyperthermia”). The concept of thermal therapy as an immune enhancer goes back to “fever” as part of a host defence against infection. Infection-related fever is a response that vertebrates have conserved for millions of years. The importance of increased body temperature is even seen in cold-blooded species like lizards or fish. Cold-blooded species seek warm external environmental temperatures to be able to utilise heat to help control bacterial viral or fungal infections. If their body temperature is not increased, infections are lethal [127,128]. Warm-blooded mammals, including man, control their own temperature by developing fever to significantly augment the immune system’s ability to control infections [129,130].

There are at least several mechanisms to explain why temperature change enhances immune therapy. Infection-incited fever acts as a “danger signal” to alert the immune effector lymphocytes of threat and cause them to search and arms them to destroy the infectious threat or more rarely a malignant tumour [131]. It is possible that the fever signal can be powerful enough to interrupt the cancer’s ability to block the effector cells.

Infection-induced fever produces immunological intermediate molecules, including powerful inflammatory cytokines such as TNFand IL-6 that externally-induced temperature elevation (hyperthermia) may or may not, depending on the temperature range of the induced hypethermia [132,133]. Although externally-applied hyperthermia differs from infection-induced fever, both natural fever and induced temperature elevation/depression (hyperthermia, cryothermia) act to increase the immune response [131]. Low-temperature hyperthermia (< 41 °C) induces vasodilation whereas high temperature hyperthermia causes initial vasoconstriction, although both temperature ranges disrupt abnormal tumour vasculature [134,135].

Inducing temperature alterations using either systemic or local thermal therapies can inhance cancer treatment [136]. What level of temperature modification is necessary is the question that needs to be researched. In the clinic, moderate or “fever-range” temperature appear to increase immune response and are used as an adjuvant therapy rather than as a “stand-alone” treatment [137]. In an in vivo model, physiological range hyperthermia (38.5–40 °C) enhances host immune defence against cancer [136,138].

Elevated temperature increases infiltration of cytotoxic CD8+, CD4 + T lymphocytes into the tumour and tumour microenvironment [139]. To explore another mechanism of how fever-range hyperthermia enhances cancer immunity, Mace et al. showed that mild hyperthermia (39.5 °C) increases the rate of antigen-specific conjugate formation with antigen presenting cells (APCs) as well as the speed of transformation of naïve CD8 + T cells to effector CD8 + T cells. Mild hyperthermia increased the rate that naïve effector cells matured to become cytotoxic effector cells so that they can penetrate the tumour and tumour stroma to kill the tumour cells [140,141]. Hyperthermia increases migration of APCs and T-lymphocytes to increase the immune response [142–144].

Notably, fever-range hyperthermia therapies are not the only temperature levels that augment immune response [145–147]. Acute changes of both extreme heat and extreme cold enhance the host immune response. Thermal ablation of tumours, a minimally invasive treatment, is a commonly used therapy for primary and metastatic liver lesions as well as other malignancies including breast, prostate, lung, renal and brain. Thermal ablation can also treat bone metastases [148–154]. The extreme temperatures used to ablate tumours increase host immune function. From ablation of small, unresectable tumours to experimental therapies for larger localised tumours, the modalities of percutaneous radiofrequency, microwave ablation and high-frequency ultrasound’s (HIFU) all increase immune function [154,155,156]. Interestingly, extremely low temperatures induced by cryotherapy also enhance the immune system’s ability to kill cancer [157–160].

An explanation of why these broadly diverse temperature ranges increase anti-tumour immune response is that, in addition to alerting the immune system, the effector CD8 + T cells must recognise specific cancer cell neoantigen(s). As described above, malignancies frequently edit their neoantigens, removing them from the cell surface to evade immune surveillance [23]. Yet, when both moderate and extreme heat such as HIFU or extreme cold such as cryotherapy ablate a tumour, the treated cells undergo apoptosis and/or necrosis and their neoantigens are released into the tumour environment making the neoantigens available to recognition and presentation by the antigen presenting cells to ready the effector lymphocytes. In addition, tumour tissue destruction by ablation causes an inflammatory reaction which itself stimulates antigen-specific cellular immunity. The inflammatory cytokines in the microenvironment around dying tumour cells attract dendritic cells (DCs) and macrophages (the APCs). The APCs then scavenge cellular debris containing tumour antigens and then proceed to lymph nodes and present processed tumour antigens to naïve effector T-cells (CD8 + Tcells), thus generating cellular immunity to the tumour-specific neoantigens [159,160]. Chemotherapy and radiation therapy also kill malignant cells by apoptosis and/or necrosis and increase the availability of tumour neoantigens [161,162]. However, while reports do show synergy between chemotherapy and immunotherapy [78,163], in fact many chemotherapy agents impair CD8 + effector lymphocyte infiltration of the tumour [164]. Importantly, ablative thermal therapy or cryotherapy increase the availability of malignant neoantigens to allow the effector T-cells to recognise the tumour without killing lymphocytes or causing the severe toxicities of chemotherapy and radiation therapy.

In addition to increasing the availability of neoantigens, a synergy of hyperthermia with immune therapy occurs because temperature change affects various other mechanisms that include depletion of regulatory T cells [165], and release of inflammatory cytokines [166]. Thermal therapy can increase efficacy of anti-check point inhibitors both by increasing the immune cell recognition of tumours and by enhancing the ability of the “educated” effector T-cells to traverse stroma, enter the tumour mass and kill the cancer cells [167].

Today we fully accept the immune-enhancing role of thermal therapy; yet, so far no one has published results of combining either checkpoint inhibitors blockers or antigen-specific adaptive transfer modalities with thermal therapy in the clinic. In part, the neglect can be explained because to time an immune therapy to combine with heat has not been studied. Since timing of any therapy with thermal therapy is critical to a positive combination treatment [138], the timing of a combination of each anti-checkpoint inhibitor or adoptive therapy needs to be worked out in vivo, prior to clinical treatments.

Abscopal effect

Remarkably, spontaneous regression of untreated distant metastases can occur after thermal ablation of the primary tumour [168,169]. This interesting phenomenon is not unique to thermal therapy. Radiation oncologists have observed for years that radiation therapy can induce regression of distant metastases outside the treatment field, although this is an infrequent occurrence. This regression of a distant tumour during radiation of another malignancy is referred to as an abscopal effect [170,171]. Host immune system mediates the abscopal effect. Observations of abscopal events after ablative thermal treatments fuel interest in increasing anti- tumour immunity by combining other treatments with focal thermal ablation [171].

Summary

In summary, response to immune therapies, points to three important components that relate to the tumour as well as to the host.

The first is whether a tumour expresses a sufficient number of tumour neoantigens. If tumours express a high number of tumour neoantigens, the response to checkpoint inhibitors is significantly improved. This has been specifically shown with anti-PD-1 and anti-PDL-1 inhibitors. Both chemotherapy and radiation therapy result in increased availability of neoantigens, as do HIFU, cryotherapy or moderate heat treatments. However, compared to cytotoxic therapies, thermal modalities induce low toxicity, do not induce lymphopenia, yet increase the accessibility of neoantigens to the host APCs. Whether lower temperature ranges also increase the number of neoantigens needs investigation. It would be interesting if thermal therapy researchers would examine the number of neoantigens in malignant tumours before and after therapy.

Secondly, response to immune therapies, especially the checkpoint inhibitors, is highly correlated to the numbers of CD8+ and CD4 + T lymphocytes in the tumour microenvironment. All types of hyperthermia appear to increase the numbers of effector lymphocytes, but are the lymphocytes penetrating the tumour microenvironment and reaching the tumour after heat treatment needs to be explored. These are measurements for pre-clinical thermal researchers to make.

Thirdly, the microbiota affects tumour response to immune therapy. While faecal transplant may be a method to increase response to immune treatments, no one has investigated whether thermal therapy can alter the microbiota. This is yet another interesting area for investigation.

The literature strongly supports thermal therapy-induced immune benefit to the host. What we already know suggests that combining the many forms of thermal therapy with immune therapy should increase the response rate of immune therapy, possibly greatly. Clearly, we greatly need pre-clinical and clinical investigations that address the combination of hyperthermia with an immune therapy, such as checkpoint inhibitors and antigen-specific adaptive lymphocyte transfer. The combination of thermal therapy with immune treatment has anticipated potential to increase response rate and progression-free survival of patients with both localised and advanced cancer

Disclosure statement

No potential conflict of interest was reported by the author.