Positron emission tomography imaging sheds new light on hypoxia and antitumor therapies

The effect of the hypoxic tumor microenvironment (TME) on the effectiveness of cancer treatments has received widespread attention. It is crucial to investigate the mechanisms by which hypoxia influences the efficacy of these treatments in order to improve the therapeutic outcomes for malignant tumors and the prognoses of patients. Positron emission tomography (PET) imaging is a non‐invasive, reproducible, and quantitative imaging technique that can visualize molecular biological changes in vivo. By utilizing specific PET probes, it is possible to both depict in vivo oxygen levels within the TME and evaluate cancer treatment effectiveness at various targets. This review summarizes the effect of hypoxia on various cancer treatments and examines the role of PET imaging in understanding the mechanisms of hypoxia during and after cancer treatments. It is anticipated that this review will provide new insights for improving tumor therapy from the hypoxia perspective and for early prediction and assessment of therapeutic efficacy via PET imaging.


| INTRODUCTION
Malignant tumors pose an imminent threat to human health, and continuous exploration and research on enhancing antitumor therapy efficacy is of great significance to improving patient prognosis and increasing overall survival rates. It is well known that the tumor microenvironment (TME) is in a hypoxic state, priming tumor cell proliferation and angiogenesis. 1 While oxygen in typical TME is highly heterogeneous and dynamic, rapidly growing vasculatures that make up the neovascular network are often abnormal, disorganized, and dysfunctional, leading to inadequate oxygen supply and further increased tumor hypoxia. 2 Both hypoxia and necrosis have been proven to be independent predictors of poor clinical prognosis, no matter the tumor stage, histological grade, or lymph node status. 3 Hypoxia has also been associated with histology instability, immunosuppression, different stages of metastasis (including invasion, migration, infiltration, and extravasation) and increased resistance to many types of antitumor therapies. 4 The hypoxia-inducible factor (HIF) pathway has been described as a major regulator of the cellular response to hypoxia. 1 HIFs are heterodimers composed of inducible α subunits (HIF-1α, HIF-2α, or HIF-3α) and constitutive β subunits (usually HIF-1β). Under normoxia, prolyl hydroxylase domain-containing enzymes require oxygen as a cofactor to hydroxylate the α subunit, leading to its degradation after von Hippel-Lindau-dependent ubiquitination or its inactivation by the factor inhibiting HIFdependent ubiquitination. 5 In contrast, under hypoxia, the α subunit is stabilized and complexed with the β subunit. 1 Subsequently, it translocates into the nucleus and transcribes genes with hypoxia-responsive element sequences in various promoters. At limited oxygen concentrations, the main targets are oncogenes involved in angiogenesis and glycolysis. 6 An increasing number of studies have shown that the hypoxic TME affects the efficacy of various antitumor therapies. [7][8][9][10][11][12][13] However, current monitoring methods to examine oxygen levels at tumor sites, such as oxygen polarographic needle electrodes or biopsies, are mostly invasive. 14 Therefore, in order to deepen our understanding and further explore the relationship between tumor hypoxic microenvironment and cancer theranostics, there is a dire need for a non-invasive and reproducible technique to monitor local TME oxygen levels, as well as a more sensitive means to reflect the changes in molecular biology during/after the treatment, to better inform the disease stage and perform timely intervention for improved therapeutic efficacy.
Positron emission tomography (PET) imaging is often used for in vivo research because of its non-invasiveness, real-time imaging, and quantitative advantages. A variety of specific PET probes are available to target desired biomarkers or receptors. For example, 18 F-fluorodeoxyglucose ( 18 F-FDG) is the most commonly used probe for glucose metabolism imaging in the clinic, and 18 F-fluoromisonidazole ( 18 F-FMISO) is developed to show tumor microenvironmental hypoxia. 15 In addition, with the development of novel PET tracers in recent years, more and more target-specific imaging have been achieved to facilitate cancer diagnosis and treatment. Based on these advantages, PET imaging plays an increasingly important role in exploring the fundamentals of tumor hypoxia and their mode of action regarding the efficacy of cancer theranostics.
In this review, we examine the impacts and theranostic values of PET imaging for hypoxia imaging, cancer therapy, and treatment monitoring ( Figure 1). We are particularly interested in the impact of hypoxia on cancer treatments such as chemotherapy, radiotherapy, immunotherapy, photodynamic therapy (PDT), sonodynamic therapy (SDT), and chemodynamic therapy (CDT) (Figures 2 and 3). We provide a fundamental understanding of hypoxia in cancer genesis, evaluation, and dynamic monitoring of therapy efficacy by a thorough study of hypoxic TME using molecular imaging methods.

ROLES IN VARIOUS ANTITUMOR THERAPIES
Hypoxia, or the low oxygen levels within a tumor, plays a crucial role in the development and progression of cancer. Accumulating studies have revealed that the hypoxic TME can lead to the development of therapy-resistant cancer cells, making it more difficult to effectively treat the tumor. Additionally, locally low oxygen levels within a tumor reduce the effect of radiation therapy. Moreover, hypoxia promotes angiogenesis, allowing the tumor to further grow and spread. In this section, we will examine the current literature on the role of hypoxia in cancer therapies and discuss the underlying effects.

| Chemotherapy
Chemotherapy is currently the most commonly employed treatment in cancer therapy, but drug resistance remains a major concern in daily clinical practice. An increasing number of studies have shown that hypoxia is one of the main causes of induced drug resistance. As such, understanding how hypoxia affects the efficacy of chemotherapy will be of special importance to improved chemotherapy. In this subsection, we will go through the following seven most recognized mechanisms of hypoxiainduced drug resistance.

| Impaired drug delivery
Tumor cells and stromal cells proliferating in a restricted TME are subjected to pressure from solid tissue components, which generate solid stress and transmit this stress through the stroma, causing vasoconstriction to limit perfusion. Thus, low tumor perfusion is often accompanied by extensive hypoxia and impaired drug delivery, resulting in lower chemotherapy efficiency and poorer prognosis. 21 Collagen and hyaluronan are important targets for compressing tumor vasculature. It was reported that the angiotensin inhibitor reduced the production of stromal collagen and hyaluronan, thereby reducing solid stress in the TME to increase perfusion, leading to improved efficiency of chemotherapy in breast and pancreatic cancer models. 22 Moreover, according to the growth characteristics of tumor cells, they can be divided into proliferating, quiescent, and necrotic cells. While proliferating cells are usually located within a few cell layers of the blood vessels, quiescent and necrotic cells are farther away from the blood vessels, usually located within a radius of 50-250 μm from the blood vessels, making it a fundamental challenge for adequate drug delivery to these cells. 23

| Reduced oxygenation
The effectiveness of many chemotherapy drugs, such as cytotoxic alkylating agents, relies on the level of oxygenation within cells. These agents transfer alkyl groups to DNA during cell division, leading to DNA strand breaks or cross-links, preventing further DNA synthesis. 24 However, under hypoxic conditions, the effectiveness of alkylating agents is reduced due to increased production of substances, such as glutathione (GSH), which may compete with the target DNA for alkylation. 8 Additionally, cytotoxic chemotherapy drugs are not as effective under hypoxic conditions due to decreased free radical production. Free radical scavengers, dehydrogenase inhibitors, and substrates prevent the formation of DNA single-strand breaks and decrease the cytotoxic effects of this class of chemotherapy drugs and their efficacy. 8

| Cell cycle arrest
Some chemotherapeutic agents affect specific phases of the cell cycle, such as alkylating agents and antimetabolites, which influence mainly the process of DNA synthesis, causing damage to the chromosome and initiating apoptosis. 25 Hypoxia slows down cell cycles and induces a pre-S-phase block of DNA synthesis in actively proliferating cancer cells. The cell-cycle protein-dependent F I G U R E 1 Graphical summary of the relationship between hypoxia and the effectiveness of various antitumor therapies.
kinase KIP1 (p27)-induced G1/S block was shown to be closely associated with hypoxia. 26 As a result, the effectiveness of these chemotherapy agents on slow-cycle tumor cells is reduced under hypoxic conditions. 8 2.1.4 | Inhibiting drug-induced apoptosis p53 is a transcription factor that regulates genes involved in controlling the cell cycle and apoptosis, maintaining homeostasis under stressed conditions, and belongs to tumor suppressor genes. 27 In normal cells, p53 can induce apoptosis. However, when cells are mutated or the p53 gene itself is mutated, it can lead to the progression of normal cells to tumor cells. 28 HIF-1 can inhibit druginduced apoptosis by regulating p53 and also causes chemotherapy resistance. Under hypoxic conditions, normal cells accumulate p53 through an HIF-1α-dependent mechanism and arrest at the G0/G1 phase checkpoint of the cell cycle through a non-p53-mediated pathway. 29 This can lead to the selective retention of tumor cells with p53 mutations, promoting a more malignant phenotype.  Metallothioneins are a family of low-molecular-weight proteins found in cells that have a high affinity for various heavy metals and are thought to contribute to platinum-based drug resistance. In many tumors, metallothioneins are upregulated by the hypoxia-activated HIF-HRE system through the extracellular regulated protein kinases (ERK)/mammalian target of rapamycin (mTOR) pathway. [32][33][34] The -SH group of metallothiosin can bind to platinum-based drugs, resulting in drug inactivation and insufficient doses of platinum-based drugs reaching tumor cell DNA. 35 Studies of tumor tissue samples from patients with hepatocellular carcinoma undergoing platinum-based chemotherapy have also confirmed that metallothionein-positive cells were significantly increased in non-responders. 36 2.1.7 | Increased drug efflux At a later stage of chemotherapy, a major presentation of chemo-resistance is multidrug resistance (MDR), which is the ability of cancer cells to develop resistance to various chemotherapeutic drugs. 37 One of the key factors contributing to MDR is drug efflux caused by Pglycoprotein (P-gp), a transmembrane protein that reduces the intracellular concentration of chemotherapeutic drugs by acting as an efflux pump. 38 Hypoxia affects the expression of the multidrug resistance 1 (MDR1) gene, which encodes the P-gp protein. In various types of cancer cells, such as glioma, gastric cancer, breast cancer, and colon cancer, high levels of HIF-1 in hypoxic TME are found with increased P-gp expression. 39

| Radiotherapy
Radiotherapy uses ionizing radiation to damage DNA and further destroy cancer cells. 40 During the process, high-energy photons from X-rays or gamma rays produce DNA free radicals (DNA%), which are rapidly oxidized by oxygen, leading to the break of DNA double strands, formation of oxidized bases, and eventually cell death. 40 Under normoxic conditions, due to the presence of sufficient oxygen, reactive oxygen species (ROS) and hydrogen peroxide are produced efficiently. 41 Besides, free radicals generated from critical targets (R·) can also react with oxygen to produce first peroxyl radicals (ROO·) and ultimately ROOH, a more stable and detrimental free radical. 42 However, under hypoxic conditions, these reactions are limited due to the lack of oxygen and instability of the free radicals, leading to more repairable damages for tumor cells. Due to the absence of oxygen, intracellular thiols, such as GSH, reduce DNA% and hinder the production of double-strand breaks, resulting in reduced damaging effects of ionizing radiation on DNA as well as the overall killing of cancer cells. 42 Deep hypoxia promotes proteomic and genomic changes, and genomic instability in turn promotes the de-differentiation and progression of aggressive cancer genes and phenotypes. 9 Gray et al. in the 1950s demonstrated that hypoxic cells required a threefold higher dose of radiation to sustain damage when compared to normal cells, highlighting the significance of hypoxia in radiation resistance. 43 In addition, hypoxia upregulates HIF-1α, which also promotes radiation resistance. 40 Taken together, hypoxic TME directly hinders radiotherapeutic efficiency by offering a reductive environment, preventing DNA damage, and increasing radiation resistance.

| Immunotherapy
Immunotherapy is one of the most promising therapeutic modalities for oncology. Currently, the main focus of tumor immunotherapy lies on immune-checkpoint inhibitors (ICIs), which targets the regulatory pathways of T cells and aims to enhance effector T cell toxicity by inhibiting the relevant immune checkpoints, thereby generating antitumor immune responses. 44 A variety of antibodybased ICIs (e.g., pembrolizumab, nivolumab, and atezolizumab) have been approved by the U.S. FDA for the combination treatment of melanoma and non-small cell lung cancer (NSCLC). 45 Immunotherapy has led to longterm remission in some patients and no recurrence for many years. However, according to research data, ICIs are effective in only 20%-40% of cancer patients, and in about a third of patients whose ICI treatments are initially effective, but tumors eventually progress or recur. 46 The main factors affecting the efficacy of immunotherapy include tumor antigens, their mutational loads, host immune-related genetic variants, tumor microenvironment, and gut microbiome. 47 Among them, the TME highly influences the development of tumors and treatment prognosis. 44,48 In recent years, an increasing amount of research has shown that, in addition to affecting tumor invasion and metastasis, tumor hypoxic microenvironment can induce immunosuppression and immune resistance by promoting various changes in microenvironmental stromal cell biology, including inhibiting the proliferation and activation of immune effector cells, promoting immunosuppressive stromal cell differentiation, reducing effectors, and increasing immunosuppressive factors. 5,10,[49][50][51][52][53][54] Here, we outline the effects of hypoxia on different immune cells with the aim to understand and identify key processes during the course of immunotherapy.

| Effector T lymphocytes
Effector T lymphocytes are the main cellular component of the adaptive immune response to tumor neoantigens. 55 Hypoxia can inhibit effector T lymphocytes function and proliferation, delay the differentiation, and promote the apoptosis. The presence of large hypoxic regions in the spleen and lymph nodes prevents CD8 + T cell activation by stabilizing HIF-1α and inhibiting T cell antigen receptor-mediated Ca 2+ signaling. 51 Induction of HIF-1αdependent cancer cells evades cytotoxic T lymphocyte (CTL)-mediated killing and reduces the CTL function. 56 HIF-1α upregulates vascular endothelial growth factor (VEGF) expression in TME, which can directly induce programmed cell death protein 1 (PD-1) upregulation and other inhibitory immune checkpoints, thereby promoting T cell exhaustion. 57 Hypoxic TME also reduces the production of cell proliferation factors, such as interferon-γ (IFN-γ) and interleukin 2 (IL-2), to delay the differentiation of effector cells. 58 Moreover, hypoxia induces autophagy of effector T cells. 59 Hypoxia-induced autophagy promotes tumor cell resistance to specific lysis through a mechanism depending on the signal transducer and the activator of transcription 3 (STAT3) phosphorylation. 60 Noman et al. showed that hypoxia-induced resistance of lung tumor to CTL-mediated lysis was associated with autophagy initiation in target cells.
Furthermore, the metabolic activity of hypoxic TME also affects effector T cell functions. Hypoxia leads to increased glycolysis, which, together with the action of proton transporters and carbonic anhydrase, increases the accumulation of lactate and adenosine in TME. 61 In turn, high levels of lactate block the targeted mTOR pathway and impair effector T cell proliferation and function. 4 Lactate accumulation leads to extracellular acidosis, while low pH has been shown to stimulate lymphocyte apoptosis. 61 In addition, free adenosine interacts with the adenosine A2A receptor (A2AR) on the surface of T cells, leading to the accumulation of cyclic adenosine monophosphate (cAMP), which also inhibits T cell proliferation and cytotoxicity. 61 2.3.2 | Natural killer cells Natural killer (NK) cells are lymphocytes that kill tumor cells and virus-infected cells non-specifically without presensitization. 62 Hypoxia can directly activate the PI3K/ mTOR signaling pathway in NK cells, leading to an increase in HIF-1α expression, and in turn, the expression of the protease ADAM10. This protease is responsible for shedding major histocompatibility complex class I chainrelated molecule A (MICA) from the surface of tumor cells. 53,63 Soluble MICA downregulates the expression of natural killer group 2 member D (NKG2D)-activated receptors on NK and T cells, leading to tumor immune escape. 63 On the other hand, hypoxia can indirectly downregulate major NK cell receptors (NK p30, NK p44, NK p46, and NKG2D), resulting in impaired NK cell killing of tumor cells. 64 Alternatively, hypoxia indirectly induced regulatory T cells infiltration, which activates the immunosuppressive cell transforming growth factor β (TGF-β), impeding local NK cell functions. 10,54

| Dendritic cells
Dendritic cells (DCs) are specialized antigen-presenting cells that capture tumor antigens via major histocompatibility complex (MHC) molecules and present them to antigen-specific T cells. In turn, activated T cells initiate specific anti-tumor immune responses. 65 However, the specific effects of hypoxia on DCs are still being debated.
Many research studies suggest that hypoxia negatively affects DCs differentiation and function. Research has demonstrated that hypoxia interferes with the migration of DCs to lymph nodes by downregulating the expression of the CC-chemokine receptor 7 (CCR7) via a HIF-1dependent mechanism. 66 In hypoxic TME, the upregulation of cytokines, such as IL-10 and VEGF, were found to inhibit DCs differentiation and maturation. 49 Immature DCs in hypoxic TME express high levels of HIF-1α while FANG ET AL. upregulating B-cell lymphoma 2/Adenovirus E1B 19-kd interacting protein 3, which mediates programmed DCs death. 67 In addition, hypoxia reduces DCs' surface marker expression, including MHC-II and co-stimulatory molecules (CD40, CD80, and CD86) and pro-inflammatory cytokines (IL-1β, IL-6, and TNF-α). These hypoxiadifferentiated DCs adopt a T h2 -stimulated phenotype and will impair T cell activity. 54,68 Lastly, hypoxic DCs have been found to secrete high levels of osteopontin, a protein known to promote the migration of tumor cells. 69 Another group of studies conclude that hypoxia promotes DCs differentiation and their maturation. In combination with lipopolysaccharide, hypoxia significantly increases the expression of conjugated molecules, boosts the synthesis of pro-inflammatory cytokines, and induces the proliferation of allogeneic lymphocytes. 70 Furthermore, the attenuated expression of inducible T cell stimulating factor and granzyme B was observed in T cells co-cultured with DCs lacking HIF-1α, suggesting a positive effect of HIF-1α on the function of DCs. 71 As a result of the conflicting findings and the ongoing debate, there is currently no clear consensus on the specific effects of hypoxia on DCs. Given the complexity of the TME, it is possible that different subpopulations of DCs may be affected differently by hypoxia. Besides, the timing and duration of hypoxia exposure, as well as the specific type of tumor, may also play a role in determining the effects of hypoxia on DCs. As such, additional studies are needed to clarify the effects of hypoxia on DCs and to provide a more comprehensive understanding of the mechanisms underlying these effects.

| Regulatory T cells
Regulatory T cells are a subgroup of CD4 + T cells that exert negative immunomodulatory effects through multiple mechanisms. 72 Hypoxia is proven to promote regulatory T cells recruitment at tumor sites. In hepatocellular carcinoma and ovarian cancer, the HIF-1α-dependent pathway upregulates CCL28 expression, which effectively recruits CCR10 + regulatory T cells to tumor sites via specific binding, thus inhibiting effector T cells and promoting tumor growth. 52,73 In basal-like breast tumors, hypoxia induces the upregulated signaling between CXCR4/CXCL12 in regulatory T cells. 74 Studies have yielded inconsistent results regarding the effects of hypoxia on regulatory T cell differentiation via its interaction with Forkhead box P3 (FOXP3), a crucial transcription factor for the development and function of regulatory T cells. On one hand, hypoxia-driven HIF-1 directly binds to the FOXP3 promoter region in CD4 + T cells and promotes FOXP3 transcription through a TGF-β-dependent mechanism, thereby enhancing regulatory T cell differentiation. 75 On the other hand, HIF-1α promotes proteasomal degradation of FOXP3, while HIF-1α binding to IL-6 promotes Th17 polarization and inhibits regulatory T cell development. 76 Therefore, different microenvironments and different cytokines lead to varied effects of HIF-1α on FOXP3 and regulatory T cell functions.
In addition, extracellular adenosine strongly inhibited the proliferation and effector functions of activated T cells. Hypoxia induces exonucleases on FOXP3 + regulatory T cells via HIF-1α-mediated CD73 and CD39, which convert extracellular ATP to adenosine and promote the accumulation of adenosine, leading to diminished T cell proliferation and effector functions. 77 Several preclinical studies have demonstrated enhanced immunotherapeutic efficacy when combined with the adenosine receptor or CD73 blockade. 78 Clinical trials targeting A2AR or CD73 are currently underway for tumor immunotherapy evaluation. 79

| Tumor-associated macrophages
Macrophages can be divided into two types, the antitumor and pro-inflammatory M1 phenotype and the pro-tumor and anti-inflammatory M2 phenotype. 80 M2 macrophages are associated with poor tumor prognostics via promoting immunosuppression, angiogenesis, tumor cell activation, and metastasis. 81 The polarization of tumor-associated macrophages (TAM) is influenced by TME. Under normoxic conditions, TAM differentiates to the M1 phenotype, while under hypoxic conditions, it shifts to the M2 phenotype to promote angiogenesis and immunosuppression. 82 Pro-tumor effects of M2 macrophages are manifested in the following four main aspects. First of all, they secrete or express angiopoietin, VEGF, VEGFR, cyclooxygenase-2, platelet-derived growth factor, fibroblast growth factor, IL-8, and CXCL8 to promote tumor angiogenesis. 67,83 Second, M2 inhibits the function of effector T cells and sustains tumor immunosuppression by producing arginase-1 and inducible NO synthase (iNOS). In addition, TAM inhibits CD4 + T cells by secreting IL-10 and TGF-β and increases regulatory T cells levels by inducing chemokines CCL5, CCL20, and CCL22 to suppress effector T cell functions. 84 Third, M2 is involved in promoting tumor cell invasion. TAM secretes matrix metalloproteinase (MMP) 2, MMP9, and macrophage migratory inhibitory factor to degrade fibrillar collagen in the extracellular matrix and to promote tumor cell invasion and metastasis. 67 TAM also secretes MMP7 in hypoxic areas to protect tumor cells from T cells and NK cells-mediated Fas ligand killing. 67 Fourth, M2 expresses immune checkpoints, including the ligands of the inhibitory receptors PD-1 and cytotoxic T-lymphocyte antigen 4 (CTLA-4), which upon activation inhibit effector T cells and NK cell functions. M2 also expresses ligands of death receptors Fas and tumor necrosis factor-related apoptosisinducing ligand, which trigger caspase-dependent cell death (apoptosis) in target cells. 84 Nevertheless, hypoxia promotes TAM differentiation towards M2 with pro-tumorigenic effects. Various factors that promote TAM to M2 differentiation are present in hypoxic TME, including prostaglandin E2, TGF-β, VEGF, IL-4, IL-6, and ROS. 50 Besides, tumor cells in hypoxic TME secrete chemokines that recruit TAM to TME and promote M2 differentiation through various mechanisms, including semaphorin-3, EMAP-II, endothelin (ET)-1, ET-2, CCL2, CCL5, CXCL12, and colony-stimulating factor-1 (CSF-1). 81,84 Anaerobic glycolysis produces large amounts of lactate, which promotes M2 conversion through multiple mechanisms, including activation of HIF-1α to upregulate VEGF expression, 85 downregulation of Atp6v0d2 expression in TAM via the mTOR pathway and inducing the expression of genes that maintain VEGF and M2 stability, such as Mrc1, Retnla, and Arg1, 86 and upregulation of arginase-1 to deplete arginine, which is necessary for T cell functions, leading to reduced effector T cell functions and promoting tumor development. 83 For TAM regulation, the main focus is to interfere with its transformation to M2 subtypes. VEGF is a relatively important factor in M2 conversion, and antiangiogenesis is expected to intervene. CSF-1R-targeted small-molecule inhibitors have also been shown to curb M2 transformation by inhibiting CSF-1 functions. 87,88 2.3.6 | Myeloid-derived suppressor cells Myeloid-derived suppressor cells (MDSCs) are a heterogeneous population of immature bone marrow cells consisting of immature macrophages, granulocytes, and DCs. 89 MDSCs can directly suppress DCs, NK cells, and T cell functions, contributing to tumor angiogenesis and metastasis. 89 At tumor sites, hypoxia recruits MDSCs to aggregate in TME. HIF-1α stimulates tumor cells to produce CCL26 and SDF1, which can interact with MDSCs surface receptors CXCR1 and CXCR4, further recruiting more MDSCs. 90,91 Also, HIF-1α induces tumor cells to overexpress ectonucleoside triphosphate diphosphohydrolase 2, an exonuclease that converts extracellular ATP to 5 0 -AMP to keep more undifferentiated MDSCs in TME. 92 On the other hand, hypoxia has been found to enhance the suppressive effects of MDSCs on effector T cells. This is thought to occur through the upregulation of the nucleotidases CD73 and CD39 on MDSCs, which leads to the accumulation of extracellular adenosine and the inhibition of effector T cell functions. 93 Additionally, HIF-1α has been found to bind to the HRE of the PD-L1 proximal promoter in MDSCs and upregulate its expression as well as bind to the HRE located at the microRNA-210 promoter to upregulate arginase-1 and iNOS, which suppresses T cell functions. 94,95 V-domain immunoglobulin suppressor of T cell activation (VISTA) is another immunomodulatory receptor, and HIF-1α binds selectively to the VISTA-associated promoter to restrain effector T cells. 96 Besides, hypoxia promotes the differentiation of MDSCs to TAM towards M2 with protumorigenic effects as discussed above. 97

| Photodynamic therapy
PDT is a rapidly developing and emerging therapeutic approach that has been clinically applied to a variety of cancer treatments, targeting apoptosis of tumor cells, and reducing damage to other healthy tissues, with minimal adverse effects and high patient tolerance. 98 Three elements are needed to perform PDT: a photosensitizer, a light source, and oxygen at the diseased site. PDT can be divided into two types, type I and type II 11 ; while most photodynamic processes are oxygen-dependent (type II), there is a small portion of PDT that can take place in a hypoxic environment (type I). 11 Here, we only focus on type II oxygen-dependent PDT.
PDT employs ROS, especially singlet oxygen ( 1 O 2 ) generated from oxygen to damage cancer cells. 99 The levels of oxygen present are critical for the efficacy of PDT. Tumor microenvironments are often hypoxic, and photodynamic reactions consume oxygen and cause vascular occlusion during treatment, which can reduce the blood supply to the tumor but also further decrease oxygen levels, leading to more severe hypoxia and limiting the effectiveness of PDT. 100,101 Therefore, increasing the oxygen level around the tumor tissue during PDT is a crucial step in improving its efficacy. Currently, there are two ways to improve the efficiency of PDT: first, by maximizing the delivery of drugs and photosensitizers to the tumor tissue for precise PDT, and second, by increasing the oxygen content of the tumor tissue and improving the concentration of ROS and 1 O 2 to enhance the efficiency of PDT. 102 The goal of drug delivery in PDT is to improve the targeting of photosensitizers, 103 while overall treatment aims to alter the hypoxia within tumors. This includes increasing tumor blood flow, producing more oxygen within the tumor, or delivering oxygen to the tumor. [103][104][105] One major solution under active investigation is to use nanotechnology for delivery of photosensitizers or drugs FANG ET AL.
-9 of 18 that improve the tumor's hypoxic environment in order to perform PDT more precisely and improve its efficacy. Our group designed a biomimetic oxygen delivery nanoprobe, in which the perfluorocarbons core can load high levels of oxygen and the cancer cell membrane coating enabled homologous tumor targeting. 103 The hypoxia of TME was proven to be improved and the PDT efficacy on xenografts was enhanced. 103

| Sonodynamic therapy
SDT is a novel treatment method that uses acoustic sensitizers and low-intensity ultrasound to produce ROS and kill malignant tumors. 106 The process involves exposing the tumor to ultrasound waves, which causes acoustic cavitation and activates the sonosensitizers from the ground state to the excited state. 106 This energy is then transferred to the surrounding oxygen to produce ROS, similar to how PDT works, but using ultrasound energy instead of light. 12 One advantage of SDT over PDT is that ultrasound can penetrate deeper into soft tissues, making it useful for treating deeper tumors. 107 However, similar to PDT, the main challenges of SDT are the delivery of sonosensitizers and altering hypoxia TME. For more details of SDT, please refer to the review by Cheng and his colleagues. 12 SDT is affected by hypoxia due to lower oxygen levels and high concentrations of GSH. 108 Low oxygen levels impact ROS and 1 O 2 production, while a high concentration of GSH reduces the killing power of ROS on tumor cells. 109 In one study, Liu and his group reported an innovative oxygen-self-produced SDT nanoplatform that uses a modified fluorocarbon-chain-mediated oxygen delivery protocol. 110 The platform delivers oxygen and sonosensitizers to tumor sites using hollow mesoporous organosilica nanoparticles. The supplied oxygen generates high levels of ROS through SDT, effectively suppressing the size of pancreatic xenografts. 110 Despite the promising results, SDT studies have thus far been limited to subcutaneous tumor models in rodents. In vivo tumor monitoring has been dependent on fluorescence imaging, which is constrained by tissue depth and fails to reflect the advantages of SDT's penetration depth. Therefore, developing integrated imaging and therapeutic sonosensitizers capable of monitoring SDT treatment outcomes in real time remains a significant challenge.

| Chemodynamic therapy
CDT is a form of in situ treatment that uses the Fenton reaction or Fenton-like reactions to generate ·OH radicals at tumor sites. 111 Essentially, iron-based nanomaterials dissolve in the weakly acidic conditions of the TME to generate iron ions and initiate the Fenton reaction, which produce excess hydrogen peroxide. This in turn produces ·OH and oxygen, triggering apoptosis and tumor suppression. 111 One of the main advantages of CDT is its endogenous activation and tumor selectivity, which ensures the safety of surrounding normal tissues. 13 The generated ·OH can directly damage cancer cells, while the generated oxygen can alleviate TME hypoxia. 13 Additionally, CDT can be used in combination with other therapeutic strategies to improve antitumor therapy. However, similar to SDT, GSH can reduce the ROS generated by Fenton and Fenton-like reactions, thus decreasing GSH concentrations and increasing ROS levels are key points to improve CDT efficiency. 112,113 Zhang and his team developed a novel approach to alleviate hypoxic TME by using biodegradable cancer cell membrane-coated mesoporous copper/manganese silicate nanospheres (mCMSNs). They catalyze the decomposition of endogenous H 2 O 2 to O 2 , thus providing oxygen to the hypoxic TME. 114 Additionally, degradation of mCMSNs triggered by GSH generates Fenton-like Cu + and Mn 2+ ions that can efficiently produce highly reactive ·OH and consume GSH. 114 With improved hypoxic TME and depleted GSH, the efficacy of CDT was enhanced to inhibit the growth of tumors in xenografts. 114

MECHANISM OF HYPOXIA AND ANTITUMOR THERAPY
Currently, the most common method for monitoring the degree of hypoxia is through invasive techniques such as oxygen polarographic needle electrodes and immunohistochemical analysis. 14 However, there is a growing need for a non-invasive method that can effectively monitor both oxygen levels and therapeutic efficacy. One potential solution is PET imaging. This non-invasive, real-time monitoring tool has gained significant attention in recent years. By designing specific probes for different targets/receptors and labeling them with radionuclides, it is possible to visualize relevant metabolic processes and non-invasively demonstrate dynamic changes of hypoxic TME in vivo (Table 1, Figures 2 and 3).

| Hypoxia PET imaging
The most commonly used PET imaging agent for hypoxia in clinical settings is 18 F-FMISO, which was first developed in 1986. 15 FMISO is a derivative of the nitroimidazole compounds, which have been proven to be effective sensitizers for hypoxic cells. The nonpharmacological doses of FMISO have been radiolabeled with 18 F as a probe, 18 F-FMISO, to image hypoxic tissues in vivo through PET. 115,141 This PET tracer has been widely used in tumor imaging, particularly in head and neck cancers, and it is used to indicate radiation or chemotherapy resistance 115 (Figure 2A). A promising second-generation nitroimidazoles is 18 F-fluoroazomycinarabinofuranoside ( 18 F-FAZA) ( Figure 2B). With higher perfusion and faster clearance from blood, 18 F-FAZA PET imaging presents a higher tumor-to-background ratio when compared to 18 F-FMISO. 116 The exogenous marker, 2-nitroimidazolpentafluoropropylacetamide (EF5), is clinically relevant and used in polarographic electrode studies to indicate tumor hypoxia. EF5 has also been applied in immunofluorescence imaging. 117 Radiolabeling EF5 with 18 F produces 18 F-EF5, which has been demonstrated to measure changes in the level of tumor hypoxia in response to the hypoxia-activated prodrug SN30000. This early response biomarker shows promise for combined radiotherapy in NSCLC xenografts as shown by Chitneni et al. 117 Similarly, other studies on patients with head-and-neck cancer have indicated the potential utility of 18 F-EF5 PET imaging in hypoxic TME. 118 The third-generation nitroimidazole analog, 18 F-flortanidazole ( 18 F-HX4), has been developed and validated through preclinical and clinical studies. Compared to 18 F-FMISO and 18 F-FAZA 119 ( Figure 2C). 18 F-HX4 exhibits more hydrophilicity and faster clearance by promoting renal clearance. 18 F-HX4 has been used to assess tumor hypoxia in NSCLC, 120 head and neck squamous cell carcinoma, 121 esophageal cancer, and pancreatic cancer. 122 In contrast, 64 Cu-labeled diacetyl-bis(N4-methylthiosemicarbazone) analogs ( 64 Cu-ATSM) are a different type of hypoxia PET tracer. They demonstrate retention under hypoxic conditions, which is dependent on the redox state of the cell ( Figure 2E). 123 However, there are studies showing that the specificity for hypoxia is dependent on the tumortype, 124 and the uptake of 64 Cu-ATSM is not always correlated with traditional hypoxic markers. 20 These findings have reduced the promise of 64 Cu-ATSM for the assessment of tumor hypoxia.

| PET imaging of DNA damage repair
Assessment of DNA damage repair enzymes may contribute to chemo-or radiotherapy evaluation. As described above, hypoxia increases the activity of DNA repair enzymes, thereby decreasing chemotherapeutic efficacy. As such, probes targeting DNA repair enzymes can dynamically monitor the level or activity of DNA repair enzymes and thus monitor and assess chemotherapeutic efficacy. The 17-member family of poly(ADP-ribose)polymerase (PARP) proteins play an essential role in several T A B L E 1 Representative targets and PET probes developed for hypoxia as well as therapeutic evaluation.

Targets
Probes References DNA damage repair pathways, especially PARP1. 142 A series of PARP1-targeted PET tracers have been investigated in both preclinical and clinical studies, in which most focus on the evaluation of PARP-targeted therapy and radiotherapy. 143 Wilson et al. designed a PARP1targeted probe 18 F-olaparib and demonstrated that radiation may well increase PARP1 expression and 18 F-olaparib uptake in tumors, and the radioactive uptake correlated linearly with PARP1 expression 125 ( Figure 3A). Based on these data, PARP-targeted PET agents offer a potent toolbox to investigate the mechanism of hypoxia and chemo-or radiotherapy resistance.

| PET imaging of P-gp
MDR is associated with increased chemotherapeutic drug efflux by P-gp, which can be upregulated by hypoxia. 38 PET probes that monitor P-gp functions in vivo can be used to explore the relationship between hypoxia and chemoresistance, deepen our understanding of its in vivo function under pathophysiological conditions, predict intracellular drug concentrations, assess its role in MDR cancers, and contribute to individualized precision therapy. While most P-gp locates on the blood-brain barrier, 144 it is also widely expressed on cell membranes of the kidney, liver, and intestines. 145 There are various PET agents that target P-gp to measure its function, some of which have been used in humans, such as 11 C-verapamil and 11 Cloperamide. 126 However, many current PET probes targeting P-gp are mainly substrates or inhibitors of P-gp, and the main challenges to overcome are achieving high selectivity for P-gp and minimal radiometabolites to ensure accurate measurement of P-gp function and to investigate whether P-gp inhibitors can improve MDR. 144 Therefore, if the relationship between hypoxia and P-gp expression and chemoresistance can be clarified, it is expected to improve chemotherapy efficacy by addressing hypoxia.

| PET imaging of oxidative stress
Radiotherapy relies mainly on ROS and ROOH to damage DNA and kill tumor cells. Thus, PET probes that measure ROS and ROOH levels can help predict the efficacy of radiotherapy. Currently, a number of PET probes are developed to target ROS. An 18 F-labeled dihydroethidium (DHE) analog, 18 F-DHE, was used in PET imaging to measure ROS levels in vivo, and it was found that DOX-treated mice with inflammation showed two-fold higher oxidation than the controls. 127 Another example is 18 F-DHMT, a novel PET agent synthesized to quantify the production of myocardial ROS in large animal models. 128 The challenge of ROS PET imaging is that ROS are small molecules with short lifetimes, making it difficult to use conventional receptor-ligand binding strategy. 146 Additionally, many ROS-targeted agents are easily oxidized during the synthesis process, which make the probe synthesis more difficult.
In light of present issues on direct imaging ROS, the downstream oxidation processes provide alternative opportunities for PET imaging at tumor sites. GSH is the predominant endogenous cellular antioxidant, upregulated by hypoxia, and limits the efficacy of radiotherapy, PDT, SDT, and CDT. 42,109,112 PET imaging targeting GSH has the potential to quantify its amount in the TME and helps patient stratification. Cysteine is the rate-limiting substrate for GSH biosynthesis. PET imaging targeting cysteine/glutamate transporter (system x c − ) is able to quantify x c − activity in vivo in response to oxidative stress in tumors. 147 Figure 3B). It is believed that with the gradual development of radiosynthesis technology, more and more probes for oxidative stress will be used for PET imaging to predict and evaluate radiotherapy, PDT, SDT, and CDT efficacies.

| PET imaging of A2ARs
The hypoxic microenvironment increases glycolysis in tumor cells, leading to the accumulation of intracellular lactate and adenosine. 148 In return, the interaction of adenosine with A2ARs on the surface of T cells inhibits T cell proliferation and function. 149 A2ARs are widely expressed in the brain, heart, lungs, and spleen, and are involved in multiple physiological and pathological processes, many of which remain unknown. 137,150 A2ARstargeted PET agents can help reveal the underlying mechanisms. There are two types of A2Ars-targeted PET agents, including xanthine-and triazolopyrimidine-based tracers, some of which have been applied in humans, such as 11 C-TMSX, 11 C-SCH442416, 11 C-preladenant, 11 C-KW6002, and 18 F-MNI-444. 143 However, most studies focused on A2ARs in the brain and explored tracer distribution in neurological disorders. Given the roles of A2ARs in inflammation, cardiovascular disease, autoimmune diseases, and cancers, [151][152][153] A2ARs-targeted PET imaging offers a brand new opportunity to explore the interaction between A2ARs and T cells to further reveal the relationship between hypoxia and immunotherapy.

| PET imaging of T cells
CD8 + T cells are effector cells that play an important role in immunotherapy, and many PET probes are available to monitor CD8 + T cell activity. A phase I clinical trial is undergoing to use of an anti-CD8 radiolabeled minibody, 89 Zr-Df-IAB22M2C, in patients with metastatic solid tumors to detect CD8 + T cells biodistribution. 138 Granzyme B is a serine protease released by cytotoxic lymphocytes during the immune response and is one of the two main mechanisms by which lymphocytes mediate cancer cell death. 154 The level of granzyme B represents the activity of cytotoxic lymphocytes and reflects the strength of the antitumor immune response. 154 Larimer et al. reported a peptide GZP that specifically targets granzyme B. After labeling with 68 Ga, the probe 68 Ga-NOTA-GZP, which specifically targets the active secreted form of granzyme B, can reflect cytotoxic lymphocyte activity 139 ( Figure 3C). Therefore, it is possible to distinguish activated and inactivated lymphocytes through PET imaging, and by combining them with tumor microenvironmental hypoxia, it is promising to dynamically explore the relationship between hypoxia and immune effector cells, quantitatively monitor and analyze the immune response before and during treatment, and predict the efficacy of ICIs.
The utilization of PET imaging in evaluating the effectiveness of anti-tumor therapy has seen significant advancements. By incorporating the analysis of tumor microenvironmental hypoxia levels with these imaging results, a deeper understanding of the correlation between hypoxia and anti-tumor therapy can be attained. This can offer new possibilities for enhancing therapeutic effectiveness, aid in the selection of effective single or combination therapies, and foster the advancement of novel therapeutic approaches or combinations.

| PROSPECTIVES AND CONCLUSION
Overall, there are many aspects of the impact and therapeutic value of PET imaging in the study of hypoxia, cancer therapy, and treatment monitoring that require further exploration. Hypoxia leads to increased glycolysis, which results in the accumulation of lactate in TME. Lactate affects the ability of effector T cells to kill cancer cells and also promotes the differentiation of TAM into pro-tumor M2 cells, reducing the effectiveness of antitumor therapies. 61 If effective PET probes targeting lactate can be developed, it would be possible to dynamically and quantitatively demonstrate the amount of lactate in the TME and quantify the degree of glycolysis, which would help to regulate the oxygen content of the TME and enhance anti-tumor efficacy. Additionally, chemotherapy resistance is associated with specific proteins, such as metallothionein, which is the major cause of cisplatin resistance. 32 PET probes that monitor metallothionein can help predict cisplatin resistance early, allowing for adjustments to treatment regimens. In regard to immunotherapy, PET probes targeting immune cell membrane receptors, immune-related factors, and stroma-associated molecules of tumor cells have the potential to monitor the function of different immune cells and predict the intensity of immune responses.
Although many hypoxia PET probes have been used in clinical settings, they are still limited by several issues, including increased uptake in background tissues due to their high lipophilicity and sluggish clearance from the body, leading to escalated radiation damage. In light of these issues, we present two potential avenues for the development of hypoxia PET probes. First, investigations indicate that certain fluorescent agents for hypoxia imaging, such as the phosphorescent dye hold promise. 155 It may be feasible to initiate with fluorescent agents and modify them into tracers appropriate for PET imaging. Second, a majority of existing hypoxia PET probes are radiolabeled with 18 F, 64 Cu, or 125 I, which encounter difficulties of availability or intricate labeling processes. A novel direction for exploration is the use of 68 Ga, which offers advantages, such as ease of acquisition, a shorter half-life, and eased toxicity concerns. In fact, the utilization of TRAP as a chelator for 68 Ga-labeled hypoxic probes is reported to enhance hydrophilicity, resulting in fast clearance and a reduction of non-essential radiation dosage. 156 In conclusion, we have summarized the intricate relationship between the hypoxic microenvironment of tumors and the various anti-tumor therapies available to us, including chemotherapy, radiotherapy, immunotherapy, PDT, SDT, and CDT. We have delved into how hypoxia can greatly impact the effectiveness of these treatments, but with the use of PET imaging, we have the opportunity to revolutionize the way we understand and combat cancer. PET imaging provides a non-invasive, quantitative, and dynamic way to monitor the hypoxia of the TME and therapeutic efficacy, giving us the ability to understand cancer at a deeper level. By targeting specific biomarkers of hypoxia using PET imaging, we can pave the way for future breakthroughs in cancer treatment. We sincerely believe that we are on the cusp of a new era FANG ET AL.
-13 of 18 in cancer research, where the potential for new treatments to be more effective and the ability to predict and monitor tumor therapeutic efficacy in real time are within our grasp.