Principles of Immuno PET

13.2.1 Introduction 

At the turn of the millennium, a landmark paper detailed six hallmarks of cancer, all of which focused on the cancer cell [14]. A decade later, a revision identified avoiding immune destruction as an emerging hallmark [15]. Targeting the immune system therapeutically has now become a major focus of academic research and the pharmaceutical industry, with many agents already in routine clinical use or in development. These treatments are fundamentally changing the prognosis of many cancers that previously had appalling outcomes once metastasis had developed. Melanoma is a case in point. The introduction of ipilimumab, an antibody against cytotoxic T-lymphocyte antigen-4 (CTLA-4), saw some remarkable responses [16], but even early in its clinical development challenged the way efficacy of such agents was assessed due to atypical response patterns [17]. The development of agents that blocked the programmed death-1 and programmed death-ligand 1 (PD1/PD-L1) immune checkpoint further advanced therapeutic options for several cancers with traditionally poor outcomes [18]. Although most widely established in the treatment of melanoma, both anti-PD-1 and anti-PD-L1 agents can produce anticancer responses in a growing list of malignancies, including lung cancer, renal carcinoma, triple-negative breast cancer, and lymphoma. Nevertheless, within these cancer types only a proportion of individuals respond, and not all responders continue to do so indefinitely.

Clinical translation of immune modulating agents was built on decades of research into the role of the innate and adaptive immune systems in cancer. Importantly, there has been a progressive shift in our understanding of cancer over the past quarter of a century or more, with increasing recognition that cancers involve a symbiosis between transformed cancer cells and a complex tumour microenvironment (TME) that involves numerous other non-malignant cells and a chemical crosstalk that supports growth, invasion, and metastasis of the malignant cells [19]. In particular, paracrine signalling between cells of the TME, which involves cytokines such as transforming growth factor beta (TGFb), and immune checkpoints, which enable cancer cells to evade T-cell-mediated immune killing, are recognized as key drivers of cancer progression [20].

In parallel with this explosion in knowledge about cancer biology, molecular imaging with positron emission tomography (PET) has adapted to this evolving therapeutic landscape by establishing new methodologies for assessing response to immunotherapy and imaging the immune microenvironment. This has involved both the conventional oncological tracer, [18F]fluorodeoxyglucose (FDG), and novel PET tracers [21–24]. In this review, the role of PET in assessing the immune TME will be discussed.

 

13.2.2 Key Components of the TME

As background, it is important to understand the key components of the TME that impact the ability of the immune system to both recognize and then kill cancer cells (Figure 1).

There are multiple cell types that are important in supporting the growth of primary and metastatic tumours. First and foremost are those that create the structural and nutritive architecture that is required for establishment of cellular aggregates of more than a small cluster of cancer cells. These include elaboration of extracellular matrix (ECM) and vascularization. Cancer-associated fibroblasts (CAFs), and endothelial cells and pericytes, respectively, are important in these processes.

The fundamental role CAFs play in cancer progression is now well recognized [25]. CAFs impact cancer invasion by remodelling the ECM, modify macrophage and endothelial function by secreting soluble factors, support tumour growth via metabolic effects such as lactate shuttling, and, importantly for this discussion, interfere with T- cell function through immune crosstalk [26,27]. The proportion of CAFs varies in different types of cancer, and even within specific cancer groupings. For example, an early subclassification of ovarian cancer [28] described a C1 subgroup that was enriched for genes associated with stromal elements, particularly myofibroblasts expressing fibroblast activation protein (FAP). This subtype demonstrated significantly higher levels of desmoplasia within the ECM, low tumour infiltration of lymphocytes and the poorest survival of all high-grade ovarian cancer genomic signatures, providing important insights on the interaction of CAFs with the immune system. FAP-high CAFs are particularly associated with regulatory T-cell (T_reg)-mediated immunosuppression.

The process of neovascularization is similarly vital for cancer progression [29]. Endothelial cells and supporting pericytes cooperate to allow angiogenesis, which is a hallmark of cancer (1). Angiogenesis occurs in a stage-dependent and type-dependent manner [30]. Unlike normal vasculature, neovascularization in tumours is structurally disorganized, leading to aberrant blood flow that limits oxygen delivery. Hypoxia is thus common within cancer deposits and is itself involved in a pro-angiogenic feedback loop, as well as driving mutational and epigenetic changes in cancer cells themselves [31]. While increasing mutations would tend to generate immunogenicity, hypoxia simultaneously impacts the immune response by attracting and metabolically reprogramming various immunosuppressive cells in the TME, including myeloid-derived suppressor cells (MDSCs) and polarization of macrophages from the M1 to M2 phenotype. Tumour blood vessels express FAS ligand (FASL), which is involved in cell death signaling, and PDL1, which inhibits lymphocyte infiltration into tumours. Sprouting endothelial cells (ECs) secrete high levels of vascular endothelial growth factor (VEGF), further increasing angiogenesis and immune suppression through recruitment and differentiation of MDSCs. Poor pericyte coverage leads to leakage of fluid into the extracellular space with an attendant increase in interstitial pressure. These factors have important effects on the number, types and function of immune cells that infiltrate the TME [32].

While growth, invasion and metastasis are important in establishing the multiple sites of cancer spread that reduce the duration and quality of life of millions worldwide [33], the complex interaction of various types of immune cell is also directly involved in the survival of cancer cells. Conceptually, the cancer immunity cycle provides a sequence of critical phases that must be overcome to allow cancer cells to avoid T cell killing. Broadly, these can be divided into initiating events that activate and recruit T cells, trafficking and infiltration of those cells into tumour sites, and the process of cell killing through recognition and attack by specialized cells [34]. At a more granular level, this cycle involves release of cancer cell antigens and presentation of those antigens to cells that activate and prime T cells, which then traffic to and infiltrate tumour deposits where recognition and killing of cancer cells occur. Failure can occur at any of these steps, and strategies to overcome immunosuppressive factors form the basis of modern immuno-oncological approaches to cancer. Anti-CTLA-4 agents primarily impact the earlier phases of the cancer immunity cycle, whereas anti-PD1/PD-L1 agents enhance the recognition and killing of cancer cells.

Mutations and epigenetic changes involved in transformation of a cell into a malignant phenotype are a central part of the oncogenic process. By creating neoantigens, tumour mutational burden contributes broadly to immunogenicity. These neoantigens, arising from non-synonymous mutations, are typically released from necrotic cells and are then taken up by antigen-presenting cells (APCs), particularly including dendritic cells (DCs), in the TME. Through secretion of regulatory factors, maturation of DCs is inhibited by immune-suppressing cells within the TME. T_reg cells are a subset of lymphocytes characterized by expression of FOXP3, with MDSCs important amongst these. Once activated, T_reg lymphocytes excrete various signals, including cytokines IL-10, IL-35 and TGFβ, which inhibit antigen presentation by APCs and suppress CD8 + T cell proliferation and function [35]. Tumour-associated macrophages (TAMs) can also suppress T cell activity through secretion of factors, including various interleukins (IL-1, IL-4, IL-6, IL-10), TGFβ and tumour necrosis factor (TNF) [36]. They also impact vascularization through secretion of VEGF. Failure of maturation of DCs impairs T cell expansion and differentiation, is mediated by multiple soluble factors, and is particularly regulated by CTLA-4. Anti-CTLA-4 agents such as ipilimumab, on the other hand, induce T cell expansion and increase the peripheral T cell recognition of neoantigens [37].

Following antigen uptake, DCs migrate to draining lymph nodes, where — in the context of major histocompatibility complex class I (MCH-1) molecules — they interact to prime CD8+ T cells. These lymphocytes then traffic to sites of cancer, where they can either accumulate or be excluded. This is a complex process involving binding of lymphocytes to the endothelium through chemokine-mediated activation of integrins and integrin-dependent adhesion, followed by transmigration into the TME. ECM can directly inhibit transmigration of lymphocytes [38] and, additionally, CAFs secrete C-X-C motif chemokine ligand 12 (CXCL12), which inhibits T infiltration into tumours. Cancers are now increasingly being defined by the number of tumour-infiltrating lymphocytes (TILs) (35,39). Inflamed tumours include a rich CD8+ T cell infiltrate, whereas the immune desert phenotype is characterized by a paucity of such cells. TIL abundance, functional state, antigen specificity and spatial distribution are all relevant to immune-mediated tumour control. Response to anti-PD1/PD-L1 immune checkpoint inhibitors is enhanced in inflamed tumours that additionally have a broad chemokine profile, a type 1 interferon (IFN) signature, and elevated expression of IFN-γ–induced genes [40]. Once within the TME, CD8+ T cells need to recognize and kill cancer cells, which requires T cell receptor (TCR) recognition of cognate peptide–MHC-I complexes on cancer cells and subsequent activation of the granule–exocytosis pathway, which is mediated by perforin and granzymes [41], and death receptor–ligand interactions [20]. The PD1/PD-L1 immune checkpoint is a key regulator of these effector steps in immune suppression of cancers and is therapeutically targeted by a wide range of anti-PD1 and anti-PD-L1 agents [42].  However, abrogation of the PD1/PD-L1, even in the presence of abundant CD8+ T cells, can be overcome by exhaustion of T cells. Natural killer (NK) cells are also an important component of the cancer cell killing apparatus.

Among the soluble factors involved in the immunity cycle, transforming growth factor-β (TGFβ) signalling, secreted by cancer cells and CAFs, is a prominent inducer of the epithelial-mesenchymal transition programme in cancer cells, which is important for invasion and metastasis. TGFβ also down-regulates MCH-1 molecules, thus limiting T cell recognition, up-regulates the inhibitory immune checkpoint ligand PD-L1, and enhances secretion of the extracellular enzyme CD73, which generates immunosuppressive adenosine in the TME. Together with prostaglandin E₂ (PGE₂) and other cytokines secreted by tumour cells, TGFβ  is involved in the induction of immunosuppressive phenotypes in myeloid cells, CAFs and neovascularization. This paracrine interplay between the cancer cells and stromal elements thus acts at multiple different levels to suppress immune destruction of cancer deposits.

 

13.2.3 Imaging of the Immune TME

The biological systems discussed above provide many potential imaging targets, both generic and specific. Generic targets that influence but do not directly reflect the immune TME include alterations in metabolism, neovascularization, and the presence of hypoxia. More specific targets are immune checkpoints, various immune-modulating cell types such as CAFs and MDSCs, and those cells directly involved in immune killing, particularly CD8+ cells as well as their biological products.

Although there has been much focus on the role of metabolic reprogramming of cancer cells to preferentially use glucose due to the Warburg effect [43], it is also increasingly recognized that various immune cells within the TME are highly dependent on glycolytic metabolism. For example, a shift towards glycolysis in TAMs leads to increased lactate production with a reduction in extracellular pH, which is associated with reduced CD8 + T cell abundance and function, thereby promoting an immunosuppressive environment [44]. Conversely, the processes of T lymphocyte expansion as part of the priming process and subsequent cell killing are also energy-dependent, with an increase in glycolytic metabolism being a manifestation of the phenomenon of pseudoprogression, which is most common with anti-CTLA-4 agents but can also be seen with anti-PD-1/PD-L1 agents [24, 45] (Figure 2).

As a modulating factor of immune function, imaging of neovascularization is feasible. Angiogenic targets that can be imaged using PET tracers include αVβ3 or αVβ5 integrins [46] and the vascular endothelial growth factor (VEGF) axis [47]. As discussed above, aberrant vessel architecture in cancer deposits leads to hypoxia. The stabilization of hypoxia-inducible factors (HIF) allows expression of a wide range of genes that allow adaptation of cells to hypoxia [31]. These include glucose transporter 1 (GLUT1), which regulates transport of glucose into cells. High FDG uptake, as measured most simply by the standardized uptake value (SUV), can thus reflect not only the proliferative status of cancers and immune cell infiltration, but also hypoxia. The presence of a hypoxia can be imaged using various tracers, particularly including [18F]fluoro-misonidazole [48] and [18F]fluoroazomycin arabinoside [49], though the predictive value of hypoxia imaging agents for response to immune checkpoint inhibitors remains to be evaluated. However, these are indirect  means of assessing the immune TME by virtue of their role in modulating immune cell function.

CAFs are more directly implicated in modulating the immune TME [50]. A key feature of CAFs is the expression of fibroblast activation protein (FAP), and several FAP tracers are entering clinical practice [51]. While high uptake of these tracers in tumour deposits may indicate a lower likelihood of response to immune checkpoint therapy, the effects of CAFs on the immune environment are most likely multifactorial [52]. The extent to which FAP tracers will complement or replace FDG remains unclear [53]. Nevertheless, like FDG, these agents have the advantage of integrating many processes that have adverse prognostic implications. By imaging both CAFs and neovascularization, a bispecific heterodimeric radiotracer targeting both FAP and αvβ3, ([68Ga]-LNC1007), which consists of FAPI-02 and a cyclic RGD peptide [54], may identify the presence of a more immunosuppressive TME. MDSCs, which play an important role in immune evasion of cancer cells, express CD11, for which a zirconium-labelled monoclonal antibody has been developed [55].

As the most important effectors of cell killing, imaging of CD8+ T cells themselves has been performed using radiolabelled minibodies [56]. Despite the promise of this approach, demonstration of the presence of CD8+ in the TME does not provide evidence of their functional status since both activated and exhausted T lymphocytes have expression of this target. The nucleoside salvage pathway, which is mediated by deoxycytidine and deoxyguanosine kinases, is up-regulated in activated T cells. A substrate for this pathway, 2’-deoxy-2’[18F]fluoro-9-b-D-arabinofuranosylguanine (AraG), is consequently preferentially taken up by these cells and can potentially differentiate between activated an exhausted T cells [57]. Similarly, agents that target lymphocyte activation gene 3 (LAG-3) have also been evaluated in preclinical models, with both peptide [58] and monoclonal antibody agents being developed [59]. A cyclic peptide against LAG-3 has also been tested in a small number of patients with non-small cell lung cancer [60]. A further target on activated T cells is T cell immunoreceptor with immunoglobulin and ITIM domain (TIGIT) [37]. A monoclonal antibody against TIGIT has been described, although uptake was not significantly higher than in a non-specific monoclonal antibody, suggesting that uptake may mainly have been related to disruption of the blood-brain barrier in the preclinical model used [61].

Cell surface receptors expressed on lymphocytes or cancer cells that are directly involved in immune recognition and killing represent attractive imaging targets, as they remain cell surface associated and can be expressed in high concentrations. Anti-CTLA-4, anti-PD1 and anti-PD-L1 antibodies have all been radiolabelled, primarily using zirconium-89 to enable sufficiently late imaging to allow blood clearance and tumour accumulation [62, 63]. Agents evaluated in both preclinical and clinical settings have recently been reviewed in detail [64]. Development of multispecific antibodies is an emerging therapeutic strategy, and in parallel with this, PET tracers have been developed to assess the biodistribution of these agents, as recently reviewed in detail [65]. A small adnectin that targets PD-L1 has also been used clinically [66], with further adaptations of this approach tested in preclinical models [67]. Being labelled with fluorine-18, these have practical advantages for diagnostic use.

An alternative to imaging T lymphocytes is targeting their secretory products involved in cell killing. For example, granzyme B is a serine protease that is released by activated CD8-positive T cells and natural killer cells to induce apoptosis in cancer cells. Imaging of the granzyme axis has been demonstrated preclinically [68] and also clinically [69].

Similarly, IFN-g is a cytokine released by a myriad of cells, including cytotoxic T lymphocytes and natural killer cells, which can either promote or inhibit inflammation. A zirconium-89-radiolabelled monoclonal antibody against IFN-g has been developed and proposed as a predictive tool for monitoring response to tumour immunotherapy based on preclinical studies [70]. IL-2 is a cytokine that is essential for the proliferation and effector function of CD8+ T cells, as well as the development of immunological memory [71]. For PET imaging of the cognate receptor of IL-2, IL-2R, the tracer N-(4-[18F]-fluorobenzoyl)IL-2 ([18F]-FB-IL2) has been evaluated in a first-time in-human trial [72].

 

13.2.4 Conclusion

There are thus many potential specific imaging approaches for imaging of the immune system (Table 1). Some have only been evaluated in preclinical models, but clinical translation of other agents is also occurring. Given the complexity of the immune TME and evolving therapeutic pathways, it remains unclear how these agents might be integrated into diagnostic evaluation of cancers. However, it is likely that they will play a complementary role to more generic assessment of the TME by FDG or FAPI agents. Different agents may play a role at primary diagnosis and for therapeutic response assessment [23]. At present, many of the agents are monoclonal antibodies labelled with zirconium-89, which pose a logistical challenge in the diagnostic setting due to the low administered activities mandated by the long half-life of the radionuclide and the need for delayed imaging. While long-axial field-of-view scanners have enabled more rapid and higher quality imaging of such agents [73], small molecule and peptide-based agents  labelled with either fluorine-18 or gallium-68 will likely be preferred in the diagnostic setting. As immune-modulating therapies become increasingly embedded in oncological care, it is likely that molecular imaging with PET will play an important role in selecting, planning and monitoring these treatments.

Figure legends

Table 1

Radiopharmaceuticals for PET Imaging of the Immune System

Target                                                                     Examples of radiopharmaceuticals

 

LAG3 lymphocyte-activating gene 3; IL-2 interleukin 2; IFN-g interferon gamma; CTLA-4 cytotoxic T-lymphocyte–associated protein 4 ; FAPI Fibroblast activation protein inhibitor; MDSCs myeloid-derived suppressor cells

 

 

Figure 1. The cancer immunity cycle has key components that involve antigen presentation, immune cell priming, trafficking of T lymphocytes to the tumour and cell killing. At each stage of the process, inhibitory factors can suppress the immune response, with CTLA-4 being the dominant checkpoint in the earlier stages and PD1/PD-L1 for later cell killing. Stromal cells including CAFs, T_reg, and MDSCs (not shown) also act to support immune evasion, as do physicochemical effects of neovascularization including hypoxia, low pH, and increased interstitial pressure.