Your browser does not support the NLM PubReader view.

Go to this page to see a list of supporting browsers.

Understanding Immune Cell Adaptation to Tumor Hypoxia for Maximized Therapeutic Efficacy of Immunotherapy: Biology and Non-invasive Imaging Application

Article information

Cancer Res Treat. 2026;58(1):26-47
Publication date (electronic) : 2025 April 29
doi : https://doi.org/10.4143/crt.2025.200
Taerim Oh1orcid_icon, Minwoo Kim1, Gi-Sue Kang1, Sung-Joon Ye2, Changhoon Choi3, Won Park3, Michael Hay4, Hiroshi Hirata,5orcid_icon, G-One Ahn,1,6orcid_icon
1College of Veterinary Medicine, Seoul National University, Seoul, Korea
2Graduate School of Convergence Science and Technology, Seoul National University, Suwon, Korea
3Department of Radiation Oncology, Samsung Medical Center, Seoul, Korea
4Auckland Cancer Society Research Centre, University of Auckland, Auckland, New Zealand
5Division of Bioengineering and Bioinformatics, Faculty of Information Science and Technology, Hokkaido University, Sapporo, Japan
6Cancer Research Institute, College of Medicine, Seoul National University, Seoul, Korea
Correspondence: G-One Ahn, College of Veterinary Medicine, Seoul National University, 1 Gwanak-ro, Gwanak-gu, Seoul 08826, Korea Tel: 82-2-880-1178 E-mail: goneahn@snu.ac.kr
Co-correspondence: Hiroshi Hirata, Division of Bioengineering and Bioinformatics, Faculty of Information Science and Technology, Hokkaido University, North 14, West 9, Kita-ku, Sapporo 060-0814, Japan Tel: 81-11-706-6762 E-mail: hhirata@ist.hokudai.ac.jp
Received 2025 February 21; Accepted 2025 April 28.

Abstract

It is extensively documented that tumor hypoxia contributes to the failure of chemotherapy and radiotherapy. Recent evidence suggests that hypoxia is also closely involved in the resistance to immunotherapy. In this review, we highlight how immune cells that are essential for the maximized immunotherapy efficacy, including cytotoxic T cells, dendritic cells, and natural killer cells, can adapt to tumor hypoxia. We then outline previous attempts targeting tumor hypoxia (for example, modulators of tumor cell oxygen consumption, perfusion modulators, hypoxia-activated prodrugs, hypoxia-inducible factor inhibitors, and hypoxia-responsive chimeric antigen receptor T cells) discussing how these approaches have resulted in an improvement of the antitumor response to immunotherapy in preclinical or clinical settings. Lastly, we review various non-invasive techniques to detect the tumor hypoxia and immune responses. We believe that an integration of the biological knowledge of immune cell adaptation to tumor hypoxia with the cutting edge non-invasive imaging technologies may ultimately allow us not only to select for patients who would benefit the most from the immunotherapy but also to monitor their responses in a real-time manner so that we can offer them an optimal personalized medicine in the clinic.

Keywords: Tumor hypoxia; Immune cells; Immunotherapy; Non-invasive imaging

Introduction

Hypoxia is defined as the condition in which oxygen availability at the tissue level is reduced thereby affecting the cellular function and organismal homeostasis [1]. At the cellular level, when approximately 7 to 8 cells cluster together, the local oxygen gradients become insufficient with the diffusion limit of the oxygen in tissues, typically around 100-200 μm [2]. Hypoxic conditions can arise in various settings including embryonic development, wound healing, and pathological conditions such as cancer, ischemia, and chronic obstructive pulmonary disease [3].

Tumor hypoxia arises as a consequence of the abnormal tumor vasculature [4]. Present in all human and animal solid tumors, hypoxia has been recognized as a major contributing factor to failure of all anticancer therapies, including chemotherapy, radiotherapy, surgery, and most importantly the latest immunotherapy [5]. Recent clinical evidence has demonstrated that tumors from the recurrent/metastatic head and neck squamous cell carcinoma (HNSCC) patients treated with anti–programmed death-1 (PD-1) antibodies have increased expression of carbonic anhydrase IX (CAIX), a well-characterized hypoxic marker [6] and decreased infiltration of CD8+ cytotoxic T cells [7]. It is extensively documented that tumor hypoxia provides an immunosuppressive environment. For example, it has been shown that tumor hypoxia decreases major histocompatibility complex (MHC) I expression through unfolded protein response-mediated autophagy in cancer cells [8]. Moreover, major immunosuppressive immune cells including tumor-associated macrophages (TAM), myeloid-derived suppressor cells (MDSC), and CD4 regulatory T cells (Treg) have been shown to migrate towards hypoxic regions in the tumors [9]. Casazza et al. [10] have elegantly demonstrated that semaphorin 3A expression in cancer cells defines TAM positioning within the tumor. Specifically, neuropilin-1, a receptor originally identified to control neuronal guidance and axonal growth, in co-operation with vascular endothelial growth factor receptor-1 on TAM has been shown to be a critical factor for their distribution within hypoxic areas of the tumors [10]. Tumor hypoxia has also been reported to increase the expression of programmed death-ligand 1 (PD-L1) on cancer cells, TAM, and dendritic cells through activation of hypoxia-inducible factor-1 (HIF-1) [11-13] thereby inhibiting activation of effector T cell by binding to PD-1 [14].

Mammalian cells can adapt to hypoxic conditions by activating HIF, the master transcription factor regulating hundreds of genes involved in angiogenesis, glycolysis, cell survival, and apoptosis [15]. HIF are heterodimeric complexes comprising an oxygen-sensitive α subunit and a constitutively expressed β subunit [16]. There are three isoforms of HIF-α, namely HIF-1α, HIF-2α, and HIF-3α, each of which has distinct, yet somewhat overlapping roles in maintaining oxygen sensing and homeostasis in cells. The α form contains a basic helix-loop-helix domain, two Per-Arnt-Sim (PAS) domains, and an oxygen-dependent degradation domain (ODD) containing an NH2-terminal transactivation domain [17]. Only HIF-1α and HIF-2α have a COOH-terminal transactivation domain while HIF-3α has been reported to exist as multiple variants, some of which lacking one or more of these domains [17]. Under normoxic conditions, HIF-α is hydroxylated at the defined prolyl residues by prolyl hydroxylase domain (PHD) enzymes (PHD1, PHD2, and PHD3 encoded by egl nine homolog [EGLN] 2, EGLN1, and EGLN3, respectively), leading to its recognition and ubiquitination by the von Hippel-Lindau tumor suppressor [18]. Hydroxylation by the 2-oxoglutarate-dependent factor inhibiting HIF (FIH, encoded by HIF1AN) makes HIF-α subunits unable to bind to their transcriptional co-activators, the histone acetyl transferases CREB-binding protein and p300, preventing the formation of a functional transcriptional complex [19]. In hypoxic conditions, the activity of HIF hydroxylase is inhibited, allowing HIF-α stabilization and subsequent nuclear translocation, where it dimerizes with HIF-β and binds hypoxia response elements (HRE) in target genes [20]. There are stimuli other than the molecular oxygen that can activate HIF. For example, reactive oxygen species (ROS), inflammatory mediators and bacterial products such as lipopolysaccharide [21], nuclear factor erythroid 2–related factor 2, an antioxidant transcription factor [22], microRNA and long non-coding RNA [19] have been shown to regulate transcription and/or translation of HIF in normoxic cancer cells.

HIF-1α is ubiquitously expressed in most of the mammalian cells and the most studied HIF isoform. In cancer cells, HIF-1 is primarily involved in regulating glycolytic metabolism by activating genes such as glucose transporter-1 (GLUT-1) and lactate dehydrogenase A (LDHA) [23]. HIF-2α is expressed rather in a tissue-specific manner, primarily in endothelial cells, kidney interstitial cells, and neural tissues [24]. Unlike HIF-1, HIF-2 has been reported to be more critical for long-term adaptation to hypoxia [25]. For example, it regulates erythropoiesis by activating erythropoietin, thereby increasing red blood cell formation to enhance oxygen transport [26]. HIF-2 also modulates iron homeostasis through genes such as transferrin and ferroportin, ensuring efficient iron utilization for hemoglobin synthesis [27]. In cancer cells, HIF-2 has been demonstrated to act as a key oncogene, driving tumor progression through the activation of cellular myelocytomatosis (c-Myc) and transforming growth factor-α in renal cell carcinoma [28]. HIF-3α, the least understood isoform, functions predominantly as a modulator of the hypoxic response [17]. It has multiple splice variants, some of which act as dominant-negative regulators by competing with HIF-1α and HIF-2α for HRE binding [29]. For instance, the IPAS (inhibitory PAS domain protein) variant lacks transactivation domains and sequesters HIF-β, attenuating the hypoxic response [30]. While HIF-3 is less involved in direct transcriptional activation [31], it has been reported to play disease-specific contexts. For instance, dysregulation of HIF-3 has been implicated in diseases such as pulmonary hypertension, where it influences vascular remodeling, and fibrotic diseases, where it modulates tissue repair [32].

In this review, we will highlight how immune cells, especially those mediating antitumor immune response can adapt to tumor hypoxia through HIF-dependent and -independent mechanisms. Understanding how these cells respond to hypoxic conditions is particularly important in order to maximize antitumor responses to immunotherapy especially given the fact that they may already reside in hypoxia conditions in unperturbed physiological conditions such as lymphoid organ where the O2 concentrations have been reported to be below 2.5% [33] compared to the physiologic pO2 (partial pressure of oxygen) in healthy tissues that typically spans between 3.3% and 7.9% [34-36].

Impact of Tumor Hypoxia on Those Immune Cells Critical for the Antitumor Immunity

1. Cytotoxic T cells

Cytotoxic CD8+ T cells are essential in mediating antitumor immune response [37]. Exhausted T cells are a subset of cytotoxic T cells that are exposed to persistent antigen presentation thereby losing polyfunctionality and renewal capacity [38]. These exhausted CD8+ cells in tumors have been shown to experience very low levels of oxygen as demonstrated by being strongly positive for the hypoxic marker pimonidazole [38-40].

Then does hypoxia cause T cell exhaustion? There are some studies suggesting that this is the case. For example, Scharping et al. [38] have reported that murine CD8+ T cells cultured in chronic (5 days) hypoxic conditions (1.5% O2) concurrently with anti-CD3/anti-CD28 stimulation has led to T cell exhaustion via HIF-1α-independent, but Blimp-1–dependent manner suppressing mitochondrial biogenesis resulting in elevated level of ROS production from the mitochondria. Interestingly, the T cell exhaustion induced by such chronic hypoxia has been shown to be an irreversible process due to persistent activation of nuclear factor of activated T cells 1 signaling [38]. Vignali et al. [41] have demonstrated that tumor hypoxia induces CD39 expression on terminally exhausted mouse T cells, where CD39 is known to hydrolyze extracellular ATP to adenosine [42] thereby eliciting immunosuppressive functions similar to Foxp3+ CD4+ Treg.

On the other hand, there are studies suggesting that acute hypoxic exposure can enhance the cytotoxic T cell function. Cunha et al. [43] have reported that acute exposure (2 hours) of low oxygen levels (1 or 5% O2) prior to CD3/CD28 activation of 3 days results in a significant increase in effector differentiation of the mouse CD8+ T cells, upregulating molecules such as the interleukin-2 (IL-2) receptor alpha subunit CD25, granzyme B, and interferon-γ via facilitating HIF-1α protein stabilization. Similarly, another study has demonstrated that peripheral blood-derived human T cells cultured in hypoxia (1% O2) for 72 hours concurrently stimulated with CD3/CD28 antibodies have exhibited an enhanced proliferation through increased HIF-1α expression and glycolytic metabolism [44]. Interestingly, the latter study has also demonstrated that the tumor-antigen expressing chimeric antigen receptor T cells (CAR T cells) delivered to neuroblastoma tumor-bearing mice proliferate more efficiently in pimonidazole-positive hypoxic areas of the tumors [44].

2. Dendritic cells

Dendritic cells provide the most important link between the innate and the acquired immune response where they present the antigenic peptides to cytotoxic CD8+ T cells and proinflammatory T helper cells activating the lymphocytes via direct cell-cell contact and proinflammatory cytokines such as IL-12 or IL-6 [45,46]. Studies have demonstrated that hypoxia exposure (1% O2 for 4 days) of immature dendritic cells derived from the human monocytes incubated with IL-4 and granulocyte macrophage colony-stimulating factor has minimal effects on the dendritic cell differentiation (in the expression of surface markers such as human leukocyte antigen [HLA]-II and CD80) [47,48] but induces more migratory phenotypes via increase in the expression of chemokine receptor genes including CC chemokine receptor (CCR) 3, CCR2, and CXC chemokine receptor 1 (CX3CR1) [47], chemotactic response to CCR2 and CXCR4 agonists [47,48], and production of CC chemokine ligand (CCL) 20, C-X-C chemokine ligand (CXCL) 1, CXCL8, and CXCL10 [48]. Similarly, Kohler et al. [49] have demonstrated that hypoxia differentially regulates chemokine and chemokine expression in the bone marrow-derived dendritic cells by increasing CCR7 expression and CCL17 and CCL22 production through a HIF-1–dependent mechanism. Bosco et al. [50] have identified triggering receptor expressed on myeloid cells-1, a member of the Ig superfamily of immune receptor and a strong amplifier of the immune responses [51] as a novel hypoxia-inducible gene in monocyte-derived dendritic cells exposed to 1% O2 for 4 days, responsible for extracellular signal-regulated kinase 1 (ERK-1), Akt, and nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor α (IκBα) phosphorylation and proinflammatory cytokine and chemokine secretion.

A transient hypoxic exposure (48 hours at 0.5% O2) to human monocyte-derived dendritic cells has been shown to dramatically increase the cell surface expression of differentiation markers including CD40, CD80, and HLA-DR isotype, as well as the production of IL-10, and proliferation of co-cultured lymphocytes [52]. Interestingly, 5% O2 exposure for 48 hours to dendritic cell progenitors has been reported to inhibit the differentiation towards plasmacytoid dendritic cells by activating HIF-1 and its downstream effector, inhibitor of DNA binding 2 [53], a critical regulator during the dendritic cell subset [54].

3. Natural killer cells

Natural killer (NK) cells are destined to remove abnormal cancer cells through the detection of major MHC class I-deficient targets expressed on various types of cancers [55]. NK cell cytolytic action towards the cancer cells are mediated by an interaction of the cancer cell surface stress proteins such as MHC class I chain-related molecule A and B or UL16 binding protein 1-6 with the natural killer group 2 member D (NKG2D) receptor on NK cells [55]. Studies have consistently reported that hypoxia reduces NK cell cytotoxicity [56-58]. Sarkar et al. [56] have demonstrated that the cytotoxicity of healthy donor-derived human NK cells gradually decreases by lowering the O2 concentrations from 21%, 1%, 0.2%, to 0%, possibly through the reduction in the surface expression of NKG2D. They have further demonstrated that hypoxia-induced reduction in the NK cytotoxicity can be reversed upon reoxygenation, although the reduced cell surface expression of NKG2D seems to be irreversible [56]. It has been suggested that hypoxia-induced reduction in NK cell killing capacity arises due to signal transducer and activator of transcription 3 activation [57] and that these hypoxia-exposed NK cells with reduced cytotoxicity have altered cellular metabolism towards glycolysis to produce ATP accompanied by dysfunctional mitochondrial morphology and increased amount of ROS (Fig. 1) [58].

Fig. 1.

A schematic diagram demonstrating how tumor hypoxia impacts the functions of CD8+ T cells, dendritic cells, and natural killer cells. CCL, CC chemokine ligand; CCR, CC chemokine receptor; IFN, interferon; MICA/B, MHC class I chain-related molecule A and B; ROS, reactive oxygen species; ULBP-1/6, UL16 binding protein 1-6.

4. Other PD-L1-expressing immune cells

PD-L1 (also known as CD274) expressed on tumor cells and/or immune cells in the tumor microenvironment interacts with PD-1 on tumor-infiltrating lymphocytes (TILs), attenuating effector T cell responses thereby allowing tumors to escape from the immune attack [59]. Studies have shown that human monocytes can express constitutive PD-L1 [60] and that TAM are the dominant source of PD-L1 in tumors [61]. Interestingly, Wang et al. [62] have reported that PD-L1 expression is significantly increased when monocytes are being rested ex vivo and further so upon monocyte-to-macrophage differentiation. They have further demonstrated that although TAM are extensively characterized to have a pro-tumoral role, PD-L1–expressing TAM are immunostimulatory with spatial preference to T cells in tumors [62]. Interestingly, Gordon et al. [63] have identified PD-1–expressing TAM and found that they are exclusively an M2 phenotype and impaired for phagocytosing PD-L1–expressing cancer cells. It is currently unknown whether tumor hypoxia regulates PD-1 expression in TAM, although it has been previously reported that it impairs the phagocytosing function of TAM. Hussain et al. [64] have demonstrated that hypoxia therefore HIF activation leads to strong induction of Fc gamma receptor IIb (FcγRIIb) expression thereby reducing the ability of TAM to phagocytose the anti-CD20 antibody–opsonized cancer cells. Importantly, FcγRIIb expression on TAM has recently been reported to be a critical mediator for the resistance of anti–PD-1 therapy. By using intravital microscopy technique, Arlauckas et al. [65] have elegantly demonstrated that anti–PD-1 monoclonal antibodies are being transferred from CD8+ T cells to TAM through FcγR at 24 hour post-treatment impairing the immunotherapy efficacy. They have further shown that the uptake of PD-1 antibodies by TAM could be prevented by using FcγRIIb/III blocking antibodies or deglycosylated anti–PD-1 antibodies, which have allowed to prolong antibody binding to T cells therefore an improved antitumor response [65]. Xun et al. [66] have identified a novel subset of TAM expressing activated leukocyte cell adhesion molecule (ALCAM) under hypoxic conditions and that these ALCAMhigh TAM are closely distributed with the exhausted T cells in the tumor.

Clinical Evidence of Tumor Hypoxia Compromising Immune Cell Infiltration and the Immunotherapy Efficacy

A number of clinical studies have reported that tumor hypoxia can significantly interfere with antitumor immune cell infiltration hence immunotherapy efficacy. Brooks et al. [67] have demonstrated by multiplex immunostaining that CD3 T cells are spatially distributed away from CAIX-positive hypoxic areas in the tumors of HNSCC patients. Similar results have also been reported by Mayer et al. [68] reporting a mutually exclusive spatial distribution between CD8 T cells and GLUT-1–positive hypoxic areas in the lymph node metastasis of melanoma patients. They have further reported a strong negative relationship between the extent of GLUT-1 expression and CD8 cytotoxic T cell infiltration into the tumors of various types including the brain, skin of the ear, gastrointestinal track, lung, and skin [68]. Zandberg et al. [7] have reported by immunofluorescent staining that tumors from HNSCC patients who have progressed after anti–PD-1 antibodies had higher CAIX-positive areas with lower CD8+ T cell infiltration. Importantly, this study has demonstrated that the tumor hypoxia can serve as a poor prognosis factor predicting poorer outcome of anti–PD-1 therapy [7]. Similar to these results, Estephan et al. [8] have reported that CD8+ T cells not only correlate inversely with the extent of cells positive for CAIX but also are significantly excluded from the hypoxic regions in tumors of colorectal cancer patients. In melanoma patients, Xiao et al. [69] have identified six distinct tumor microenvironment archetypes based on the multicellular compositions in the tumor tissues using multiplexed imaging mass cytometry and found that responders to anti–PD-1 immunotherapy had more diversified cell type compositions with lower CAIX-positive hypoxic signals than non-responders in the invasive margin regions of the tumors. Vos et al. [70] have demonstrated that tumors from HNSCC patients with a major pathological response to nivolumab (anti–PD-1 antibodies) or nivolumab plus a single dose of ipilimumab (anti–cytotoxic T-lymphocyte antigen 4 [CTLA-4] antibodies) had a significant decrease in hypoxic signal, measured by [18F]-HX4, a positron emission tomography (PET) imaging tracer to image tumor hypoxia with an improved signal-to-background ratio than [18F]-MISO and [18F]-AZA [71], along with reduction in hypoxic RNA signature.

As there are no biomarkers derived from tumor hypoxia available, researchers have analyzed various hypoxia-relevant genes from big database such as The Cancer Genome Atlas to apply whether the gene expression can serve as prognostication and prediction of response to immunotherapy (Table 1). By incorporating various analysis techniques some of which even incorporate machine learning approach (Table 1) [72], many studies have reported that their hypoxia-relevant genes are capable of predicting the immune landscape within the tumor microenvironment as well as serving as a prognostic tool for immunotherapy efficacy in various types of tumors (Table 1) [72-83]. There are a couple of studies, however, reporting somewhat contradicting findings such that patients with high expression in hypoxia-relevant genes have displayed an increased infiltration of antitumor immune cells including the effector T cells (Table 1) [84-86].

Studies analyzing hypoxia-relevant genes as a biomarker predicting the tumor response to immunotherapy in the patients with various types of cancers

Agents Targeting Tumor Hypoxia to Improve Immune Responses

Based on the above evidence demonstrating how tumor hypoxia negatively impacts immunotherapy efficacy, we have searched for studies combining hypoxia-targeting strategies with immunotherapy. Although there are some studies that have not attempted initially to inhibit hypoxia in order to bring the synergistic effects with immunotherapy, we found that many, if not all, studies have demonstrated a positive outcome from the combination of the two modalities (Fig. 2).

Fig. 2.

A diagram demonstrating how various approaches inhibiting tumor hypoxia could improve immunotherapy efficacy. CAR T cells, chimeric antigen receptor T cells; CTLA-4, cytotoxic T-lymphocyte antigen 4; HAP, hypoxia-activated prodrugs; HIF, hypoxia-inducible factor; HRE, hypoxia response elements; NK, natural killer; ODD, oxygen-dependent degradation domain; PD-L1, programmed death-ligand 1; PD-1, programmed death 1.

1. Modulators of tumor cell oxygen consumption

Metformin is a broadly prescribed type II diabetes treatment efficiently lowering the blood glucose and insulin levels by inhibiting the liver gluconeogenesis [87]. The primary mechanism of metformin is considered to be a direct inhibition of complex 1 activity of the mitochondrial electron transport chain [87]. Zannella et al. [87] have reported that metformin significantly decreases tumor hypoxia by inhibiting tumor cell oxygen consumption and this enables a significant improvement of the tumor response to ionizing radiation when given 30 minutes prior to radiotherapy. Scharping et al. [88] have demonstrated that metformin leads to synergistic antitumor activities in B16 and MC38 tumors when combined with anti–PD-1 antibodies by reprogramming the hypoxic tumor microenvironment. They have further demonstrated that although metformin alone can increase in the number of CD44hi activated T cells, metformin itself is inefficient to abrogate the tumor growth in mice [88]. Interestingly, in that study metformin alone has also shown to increase the number of PD-1+ and Tim-3+ exhausted T cells in tumors [88] and this might have counter affected the cytotoxic activities of CD44hi activated T cells in exerting the tumor control.

Respiratory hyperoxia (60% O2) mimicking protocols of supplemental oxygen delivery to humans have been utilized as an anti-adenosinergic treatment and shown that hyperoxia itself (without combination with any immune checkpoint blockades) is sufficient to promote regression of tail-vein injected MCA205 and B16 melanoma as well as spontaneous lung metastasis from orthotopically implanted 4T1 tumors in mice [89]. Mechanistically, it has been shown that respiratory hyperoxia increases CD8+ T cell infiltration and function (interferon-γ production) while decreasing CD4+CD25+Foxp3+ Treg infiltration and their CTLA-4 expression in the lung tumor microenvironment [89]. As expected, respiratory hyperoxia combined with immunotherapy (anti–CTLA-4/PD-1 antibodies) has resulted in a significant improvement in the regression of tumor seeding models [89].

2. Perfusion modulators

Losartan, an angiotensin II receptor inhibitor, has been shown to decrease tumor hypoxia by alleviating the tumor interstitial pressure exerted by collagen I production from cancer-associated fibroblasts [90]. Datta et al. [91] have recently demonstrated that losartan normalizes the tumor vasculature, improves perfusion, and decreases tumor hypoxia and immunosuppression, allowing enhanced effector T cell function and antitumor response to anti–PD-1 antibodies.

Telmisartan is another angiotensin II type 1 receptor blocker with longer plasma half-life, higher lipophilicity, and higher binding affinity for the receptor than losartan [92]. Wadsworth et al. [93] have reported similar findings to losartan by showing that it can increase tumor perfusion by reducing cancer-associated fibroblasts and fibrillary type 1 collagen deposition in WiDr tumor xenograft in mice. Therefore, it is feasible that telmisartan may also offer a synergistic antitumor efficacy when combined with immunotherapy, although not yet reported.

3. Hypoxia-activated prodrugs

Hypoxia-activated prodrugs (HAP) are prodrugs (inactive compounds) that can be converted to active cytotoxins under hypoxic conditions capable of killing hypoxic cancer cells. The hypoxia-selective activation is achieved by one-electron reductases, usually P450 oxidoreductase [94], yielding a one-electron intermediate, which is sensitive to back oxidation in the presence of molecular oxygen [95]. Tirapazamine (TPZ), the first member of the benzotriazine di-N-oxide class is the most studied HAP, and had been evaluated up to phase III clinical trials [95]. Although there are no studies to date investigating the effect of TPZ on immunotherapy antitumor efficacy, Wang et al. [96] have observed immunogenic cell death in tumors evoking the development of tumor-specific cytotoxic T cells when an albumin-based nanoplatform containing TPZ is irradiated at near-infrared frequency. Interestingly, phase II clinical trials (NCT03259867) are currently undertaking to evaluate efficacy of trans-arterial tirapazamine embolization treatment of liver cancer followed by PD-1 checkpoint inhibitor (nivolumab or pembrolizumab) in hepatocellular carcinoma, metastatic colorectal cancer, metastatic gastric cancer, and advanced non-small cell lung cancer patients.

Evofosfamide (TH-302), a 2-nitroimidazole compound with a DNA-crosslinking phosphoramidate mustard [97], is another extensively characterized HAP, which has recently completed phase III clinical trials in pancreatic adenocarcinoma and soft tissue sarcoma (NCT01746979 and NCT01440088) demonstrating no significant benefit in increasing the overall survival of the treated patients [98,99]. Jayaprakash et al. [100] have reported that TH-302 in combination with anti–CTLA-4/PD-1 antibodies lead to a significantly improved antitumor response using the transgenic adenocarcinoma of mouse prostate–clone 2 prostate tumor model in mice. They have demonstrated that the decrease in tumor hypoxia caused by TH-302 leads to increased CD8+ T cell infiltration and decreased MDSC influx [100].

Other HAP such as tarloxotinib (a tyrosine kinase inhibitor prodrug) and CP-506 (the second-generation analogue of nitrogen mustard prodrug PR-104) have also been reported to offer synergistic antitumor activities when combined with immune checkpoint blockades through reduction in tumor hypoxia and subsequent increase in CD8+ T cells activation [97].

4. HIF inhibitors

32-134D (4-(6-bromo-1H-indol-3-yl)-2-(7-bromo-1H-indol-3-yl)thiazole) is a small molecule inhibitor of both HIF-1 and HIF-2 demonstrating potent antitumor activities in hepatocellular carcinoma model leading to alteration in the gene expression, reduced infiltration of TAM and MDSC, and increased CD8+ T cell infiltration in tumors [101]. Combining 32-134D with anti–PD-1 antibodies has resulted in a significant abrogation of tumor growth although 32-134D or anti–PD-1 antibodies themselves also offered very potent antitumor activities in that model [101].

PX-478 (S-2-amino-3-[4’-N,N,-bis(chloroethyl)amino]phenyl propionic acid N-oxide dihydrochloride), a selective HIF-1α inhibitor [102]. Luo et al. [103] have reported that PX-478 combined with anti–PD-1 antibodies leads to a significant tumor growth inhibition in Lewis lung carcinoma by inhibiting lysyl oxidase-like-2 signaling pathway reverting the mesenchymal phenotype of lung cancer cells to more an epithelial phenotype.

Many anticancer agents including imatinib, gefitinib, erlotinib, cetuximab, trastuzumab, everolimus, topotecan, taxotere, 17-AAG, YC-1, echinomycin, bortezomib, and trichostatin A have been shown to inhibit HIF-1 activity through a reduction in HIF-1α protein levels, HIF-1 DNA binding activity, or HIF-1-mediated transactivation of target genes [104]. Indeed, combination of imatinib [105], gefitinib [106], erlotinib [106], and taxotere [107] with immune checkpoint blockades have demonstrated synergistic antitumor activities in preclinical or clinical studies. Cardiac glycosides such as digoxin have also been identified as potent inhibitors of HIF-1 activity through inhibition of HIF-1α translation [108]. A recent study by Wang et al. [109] has shown a synergistic antitumor activity when digoxin is combined with a PD-1 inhibitor in 4T1 breast tumor model, although the proposed mechanism does not seem to involve HIF-1 inhibition.

5. Hypoxia-responsive CAR T cells

To overcome the detrimental effects of hypoxia on T cell activation, hypoxia-sensing or hypoxia-inducible CAR T cells have been developed. In this strategy, the expression of CAR genes can only be activated under hypoxic conditions under the control of hypoxic sensors such as the ODD [110,111] and HRE [110,112]. Examples include hypoxia-inducible CAR T cells (HiCAR-T cells) driven by HRE bearing HIF-1α ODD and the tumor antigen such as CD19, AXL receptor tyrosine kinase (AXL), and human epidermal growth factor receptor 2 (HER2) [113]; HypoxiCAR bearing dual sensors, namely C-terminal 203-amino-acid ODD onto the CAR while modifying the CAR’s promoter in the long terminal repeat enhancer region of the vector containing 9 times HRE with ErbB tumor antigen [110]; hypoxia-inducible transcription amplification CAR T cells where transactivators of the gene of interests can be regulated by both HRE and ODD with HER2 tumor antigen [111]; and hypoxia-responsive CAR T cells (5H1P-CEA CAR) bearing 5 times vascular endothelial growth factor HRE and ODD [112]. These CAR T cells all have consistently demonstrated hypoxia (1% O2)-specific presentation of CAR molecules and cytotoxicity [110-113], bypassing the systemic and lethal toxicity [110], and durable antitumor response [110-113]. Although these hypoxia-responsive CAR T cells are a very attractive strategy to target the tumor hypoxia, it has not yet been tested for their antitumor activities in combination with immune checkpoint blockades.

Clinical Trials on Hypoxia-Targeting Agents in Combination with Immunotherapy

Based on the promising preclinical results of TH-302 described above, a clinical trial of TH-302 in combination with ipilimumab (anti-CTLA4) has been conducted in 2017 in patients with metastatic or locally advanced pancreatic cancer, human papillomavirus-negative HNSCC, immune checkpoint blockade-refractory melanoma, or castration-resistant prostate cancer (NCT03098160). The results have revealed 16.7% partial response and 66.7% stable disease in 15 patients having measurable disease at baseline out of 18 enrolled patients with no dose-limiting toxicities [114]. Analysis of peripheral blood from these patients before, during, and after the treatment has shown a significant increase in CD4+ and CD8+ T cell proliferation and fewer accumulation of PD-1+LAG-3+ exhausted T cells in responders [114]. In paired tumor biopsies obtained before and at 7 weeks after the treatment have revealed that non-responding patients had increased expression of HIF-1α downstream genes and hyper-metabolic phenotype of T cells [114].

Vorinostat (suberoylanilide hydroxamic acid) is a well-known histone deacetylase inhibitor capable of decreasing HIF-1α protein expression [115] and nuclear translocation [116]. The results of phase II open label trial of vorinostat in combination with pembrolizumab (PD-1 antibodies) in patients with recurrent metastatic HNSCC and salivary gland cancer (NCT02538510) have demonstrated 32% partial response and 20% stable disease in HSNCC patients, and 16% partial response and 56% of stable disease in salivary gland cancer patients [117]. Although the study has suggested that overall survival and progression-free survival in HNSCC patients are superior, the toxicity associated with the combination has also shown to be significantly higher than pembrolizumab alone [117].

Developing Non-invasive Technologies to Detect Hypoxia and Immune Responses

1. Non-invasive imaging techniques for tumor hypoxia

In this section, we outline non-invasive imaging technologies for hypoxia in solid tumors in small animals and human patients. Some of the imaging techniques described here are used only for preclinical studies. However, these emerging imaging technologies are helpful in animal experiments for malignant tumors and immunotherapy. Therefore, we do not exclude imaging techniques which are only used in the laboratory. The pros and cons of imaging techniques will be discussed later.

1) Fluorescent and phosphorescent imaging

Fluorescent and phosphorescent imaging techniques are widely used in microscopy. Oxygen-sensitive fluorescent/phosphorescent agents are essential for this imaging modality. Some specially designed oxygen-sensitive imaging agents have been used for mapping the partial pressure of oxygen (pO2) [118-120]. Since this imaging modality is based on optical techniques, high-resolution imaging can be used to obtain biological tissue samples. In contrast, absorption and scattering of light in biological tissues hinder optical-based imaging techniques. These drawbacks are challenges for high-resolution tissue imaging in depth. Therefore, high-resolution three-dimensional (3D) fluorescent/phosphorescent imaging is generally challenging for a whole solid tumor.

2) Blood-oxygen level-dependent magnetic resonance imaging

Blood-oxygen level-dependent magnetic resonance imaging (BOLD-MRI) detects the difference in magnetic properties between oxygenated and deoxygenated hemoglobin (oxyhemoglobin and deoxyhemoglobin) [121]. It is widely applied to functional MRI for brain activity studies involving a shift in cerebral blood flow [122]. BOLD-MRI can also be applied to solid tumors to monitor tissue oxygenation [123,124]. However, BOLD-MRI does not quantitatively provide the pO2.

3) Dynamic contrast-enhanced MRI

Dynamic contrast-enhanced (DCE)–MRI is based on a series of MRI scans after rapid intravenous administration of a contrast agent such as a Gd-based contrast agent. The image acquisition sequence in this MRI technique should be fast enough to obtain the dynamic change of contrast-enhanced images like a movie. The obtained dynamic profile of the MRI scans reflects tumor vascularity and the tumor perfusion status [125]. The perfusion parameters obtained by DCE-MRI help predict the tumor treatment response [126].

4) 18F-fluoromisonidazole PET

PET is widely used for patient diagnosis in a clinical setting [127]. PET imaging with the hypoxia tracer, 18F-fluoromisonidazole (18F-FMISO), has been used to image hypoxia in cancer patients [128]. This tracer, 18F-FMISO, accumulates in hypoxic cells and can be visualized with a PET scan. The images of 18F-FMISO PET have been used to stratify patients into two groups of high and low hypoxia but, 18F-FMISO PET does not provide information on pO2 values and tissue oxygenation above the threshold of cellular uptake of hypoxia tracers.

5) Near-infrared spectroscopy and imaging

Near-infrared light has an advantage for tissue penetration compared to visible light or light with a shorter wavelength [129]. Oxyhemoglobin and deoxyhemoglobin have different absorption characteristics that can be detected by near-infrared spectroscopy (NIRS). Therefore, this spectroscopic technique detects oxygen saturation of hemoglobin in tissues [130].

6) Electron paramagnetic resonance imaging

Electron paramagnetic resonance (EPR) is a magnetic resonance technique for detecting unpaired electrons instead of the nuclei that nuclear magnetic resonance detects. There are two protocols, continuous-wave and pulsed operations, in EPR imaging. The EPR spectrum of oxygen-sensitive imaging agents (stable free radicals) can respond to the pO2 in the environment around the imaging agents [131]. Using pulsed EPR, the relaxation process of unpaired electrons responding to the pO2 can be measured [132]. Therefore, EPR imaging can provide pO2 maps in solid tumors [133]. Fig. 3 illustrates in vivo pO2 mapping of mouse xenograft models of pancreatic ductal adenocarcinoma cell lines (A, Hs766t; B, MIA PaCa-2; C, SU.86.86) and anatomic maps with MRI. Moreover, this figure shows the pO2 response to intravenous injection of pyruvate. This imaging modality is powerful for mapping the solid tumors of small animals. EPR imaging does not apply to a clinical setting because of the barrier of injecting free radical imaging agents into a patient (approval from the national regulatory agency is required). In preclinical studies, EPR pO2 mapping can provide a reference truth for 18F-FMISO PET for solid tumors [134].

Fig. 3.

In vivo pO2 mapping of mouse xenograft models of pancreatic ductal adenocarcinoma cell lines (A, Hs766t; B, MIA PaCa-2; C, SU.86.86) and anatomic maps with magnetic resonance imaging. Also, pO2 maps after intravenous injection of pyruvate are given at the time points until 60 minutes. Reprinted from Wojtkowiak et al. (2015), Pyruvate sensitizes pancreatic tumors to hypoxia-activated prodrug TH-302. Cancer Metab., volume 3, article no. 2, licensed under Creative Commons Attribution (CC BY 4.0) license [133].

7) Photoacoustic imaging

Photoacoustic imaging (PAI) uses an imaging agent to sense specific physical quantities in biological tissues [135]. The principle of PAI is the measurement of acoustic waves in biological tissues after irradiating light (laser pulse) into the tissues. PAI agent generates mechanical pressure due to heating by the reaction to irradiated light, leading to acoustic waves. As mentioned in fluorescent/phosphorescent imaging, light has limited penetration in tissues. While PAI uses light irradiation, the signals are received as acoustic waves, such as ultrasonic echo, that can propagate easily in tissue. Therefore, PAI can detect the information for the tissue in depth. In this context, a PAI agent is essential for detecting oxygenation (blood oxygen saturation and pO2) in solid tumors [136].

2. Imaging studies for hypoxia and immune response

The key mediator of hypoxia signaling, HIF-1α, regulates gene expression that influences tumor immunity under hypoxia [137]. Therefore, imaging hypoxia or oxygen-related parameters in tumors is essential information regarding immune response. As mentioned above, the non-invasive imaging techniques for mapping hypoxia and tissue oxygenation-related parameters are available for cancer studies. We briefly describe the studies of immune response and hypoxia imaging.

Using 18F-FMISO PET, hypoxia severity was monitored during checkpoint blockade [138]. This study identified the maximum tumor 18F-FMISO uptake as a predictor of immunotherapy response to the treatment of evofosfamide, a hypoxia-activated prodrug. The treatment outcome of evofosfamide was limited; however, it was enhanced by evofosfamide combined with immune checkpoint inhibitors [139].

Another study reported a longitudinal correlation between hypoxia and response to immunotherapy using 18F-FMISO PET [140]. In this study, tumor hypoxia was quantified in vivo before or during PD-1 and CTLA-4 checkpoint blockade in mouse models of breast and colon cancer. This work demonstrated that hypoxia imaging is an early predictive biomarker of the response to immune checkpoint therapy.

The studies above used 18F-FMISO PET to monitor tumor hypoxia. However, other imaging techniques, such as EPR imaging and PAI, can also be used in studies on hypoxia and immune response. The field of precision medicine needs to explore the relationship between hypoxia and immunotherapy. Such knowledge can support the effective selection of patients responding well to immunotherapy and the combination of chemotherapy and immune checkpoint blockade.

3. Pros and cons of hypoxia imaging techniques

This section discusses the pros and cons of the imaging techniques mentioned above. The three points below are essential perspectives for cancer researchers. Moreover, Table 2 [121,132,140-144] summarizes the imaging techniques mentioned above to compare their feasibility in clinical use, measurement resolution (or “sensitivity and specificity”), and limitations. Diagnosis of malignant tumors is an essential step in clinical practice. Therefore, the sensitivity and specificity of the imaging techniques are critical for cancer diagnosis. However, reading a quantitative oxygen parameter is essential for tumor hypoxia imaging techniques because quantitative hypoxia measurement such as pO2 can be a companion diagnostic test and has much information for immunotherapy, instead of an on-off image like fluorescent hypoxia cell mapping.

Accuracy, sensitivity, specificity, feasibility in the clinical use, and the limitations of non-invasive imaging techniques for tumor hypoxia

1) Ease of operation and access to the instruments

Fluorescent/phosphorescent and NIRS imaging instruments are relatively easy to access. In contrast, MRI- or PET-based imaging (BOLD-MRI, DCE-MRI, and 18F-FMISO PET) require large and expensive scanners. However, MRI and PET scanners are already working in many hospitals. Access to EPR imaging is limited at present, compared to MRI and PET scanners. PAI is more accessible than EPR imaging, but limited compared to currently available clinical imaging instruments.

2) Quantitative pO2 measurements

Oxygen-related imaging techniques do not necessarily provide quantitative pO2 values. Fluorescent, EPR, and PAI can provide tissue pO2 in solid tumors. BOLD-MRI, DCE-MRI, 18F-FMISO PET, and NIRS provide information about oxygenation using various parameters. This comparison does not criticize these imaging techniques. The effectiveness of the imaging techniques depends on their application and the demands of the studies. Researchers should select an appropriate imaging technique for their studies.

3) Clinical use

The most influential factor for clinical applications of tissue oxygenation imaging is the availability of the instruments and techniques. In this context, MRI and PET scanners are widely accessible in hospitals. Fluorescent/phosphorescent imaging and NIRS are also applicable to patients in a clinical setting. However, fluorescent/phosphorescent imaging agents should be approved for clinical applications. Currently, EPR imaging for pO2 mapping is limited to small animal experiments.

From the viewpoint of routine clinical use, the feasibility of implementing the imaging methods is discussed below. Fluorescent/phosphorescent imaging is feasible for superficial tumor regions in clinical use. Also, one can implement an optical imaging setup within a flexible endoscope. That endoscopic approach partly solves the limitation of optical imaging in depth. The availability of fluorescent/phosphorescent imaging agents for sensing pO2 is essential for routine clinical use from the regulatory viewpoint. BOLD-MRI is feasible for routine clinical use. MRI scanners up to the magnetic field of 3T are commonly used in clinical settings. Many BOLD-MRI studies for humans were performed with 3T MRI scanners. However, high-field MRI at 7T or even higher fields has been used for preclinical BOLD-MRI studies for small rodents or bigger animals such as macaque monkeys, common in neuroscience studies [145]. DCE-MRI and 18F-MISO PET are available in clinical use. Gadolinium-based contrast agents used for DCE-MRI are clinically available because standard routine 1H-MRI scans use it. Since PET scans use radioactive tracers, medical service providers should control radiation exposure to patients under national regulations. NIRS and its imaging are feasible for tumor diagnosis in clinical use [146]. EPR imaging is a powerful modality of pO2 monitoring and mapping in tumors. However, a clinical EPR imager for human patients is currently unavailable. This EPR imaging modality is mainly used for preclinical small animal studies. Although not EPR imaging, EPR spectroscopy is being used in a clinical trial for monitoring pO2 in superficial malignant tumors using implantable oxygen sensors [147]. PAI is an emerging method in biomedical imaging modalities [148]. Since PAI is feasible for clinical use, researchers should explore clinical applications more for tumor oxygenation monitoring.

4. Future perspectives for hypoxia imaging applications

1) Artificial intelligence applications

Artificial intelligence (AI) applications in medical imaging are rapidly growing and become more practical in preclinical and clinical imaging studies. Two examples of hypoxia-related applications are briefly introduced here. These AI applications will be translated into clinical applications in the future.

Due to technology developments in AI, deep learning denoising was demonstrated for 3D pO2 maps for mouse tumors in vivo with pulsed EPR imaging modality, as well as those signal amplitude maps [149]. This deep learning denoising technique provided a better signal-to-noise ratio and leading to precise pO2 analysis for tumors. Such an analysis is essential to identify hypoxic region(s) in solid tumors. AI applications for tumor pO2 mapping are emerging but promising, like other AI applications in imaging modalities.

Another application of AI is to identify tumor hypoxia using multiparametric data of MRI maps (T2, T2*, diffusion-weighted imaging, and DCE imaging) [150]. The hypoxic region predicted by deep learning was validated with pimonidazole-stained histology slice of patient-derived xenograft (PDX) rhabdomyosarcoma models and radiation-induced fibrosarcoma (RIF-1) tumors. The AI-predicted hypoxic fractions using multiparametric MRI maps were correlated with true hypoxic fractions using pimonidazole-stained histology.

2) Personalized hypoxia-targeting strategies

As described above, hypoxia imaging is beneficial to understanding the characteristics and malignancy of solid tumors. It also gives helpful information for a personalized treatment plan with a hypoxia-targeting strategy. First, the selection of patients for HAP is a critical application to enhance the outcomes of treatment [151]. Second, hypoxic mapping information can guide the radiation treatment planning (boosted radiation to the hypoxic region) and demonstrate the improved treatment outcome in preclinical studies since tumor hypoxia shows the resistance to radiation therapy [152]. PAI can also contribute personalized treatment. A study using PAI demonstrated the spatial correlation between pretreatment oxygen distribution and radiation therapy efficacy for triple-negative breast cancer PDX [153]. Like these hypoxia-targeting strategies, we should investigate the correlation between hypoxia mapping and immune response more to improve the outcome of immunotherapy.

Conclusion

Currently, there are considerable efforts to transform immunologically ‘cold’ tumors (defined as those having low levels of TIL) to ‘hot’ tumors (infiltrated with high levels of CD8+ T cells) [154]. For example, immunotherapy has been combined with radiotherapy or chemotherapy to increase immunologic cell death thereby facilitating the release of damage-associated molecular patterns to be able to recruit immune cells at higher efficiency [154]. Other strategies such as oncolytic viruses, cancer vaccines, cytokines, co-stimulatory receptor agonist, cyclic GMP-AMP synthase–stimulator of interferon genes (cGAS–STING) agonists have also been proposed [154]. However, immune dynamics in the ‘cold’ tumors, or those in transition from ‘cold’ to ‘hot’ during immune checkpoint blockade, remain poorly understood. Therefore, it is critical to develop non-invasive technologies that allow close monitoring of the tumor microenvironment (especially TIL) before, during, and after the immune checkpoint blockade and any adjuvant therapies to the immunotherapy. In a recent study, Lu et al. [155] have monitored CD4+ and CD8+ T cells in the course of anti–PD-1 antibodies, anti-CTLA4 antibodies, or combination of anti–PD-1 and CTLA-4 antibodies in 4T1 breast cancer model using repetitive PET imaging. They have observed increase in CD8-specific PET signal within 6 days and CD4-specific PET signal in 2 days in tumors of responders [155], indicating the need of serial measurements of PET imaging. Serial measurement of PET imaging analyses may not be practical in the clinic due to a number of limitations including the safety issue (repetitive exposure to radiation) as well as some technical challenges to measure and the cost of measurement as suggested by Dewhirst et al. [156]. However, immune response is a dynamic process that can be dramatically changed within minutes, hours, and days. Furthermore, it has been shown that tumor hypoxia can drastically change from day to day in some cancer patients [157]. Merging with immune cell dynamics and hypoxia using the technologies that we have outlined in this review may allow us to uncover many exciting therapeutics and their mechanisms overcoming tumor hypoxia and promoting the antitumor responses to immunotherapy.

Notes

Author Contributions

Conceived and designed the analysis: Ahn GO, Hirata H, Hay M, Park W, Choi C, Ye SJ.

Collected the data: Ahn GO, Hirata H, Hay M, Oh T, Kim M, Kang GS.

Contributed data or analysis tools: Ahn GO, Hay M, Hirata H, Oh T, Kim M.

Performed the analysis: Ahn GO, Hirata H, Hay M, Oh T, Kim M.

Wrote the paper: Ahn GO, Hirata H, Oh T.

Conflicts of Interest

Conflict of interest relevant to this article was not reported.

Funding

This study was supported by the Research Grant from Seoul National University (Multidisciplinary Research Grant, no. 550-20230098 to G.O.A.), grants from the Ministry of Science and ICT (Information and Communications Technology), Korea (grant no. RS-2024-00391742 and 2023-00218623, and 2025-02316610 to G.O. Ahn), the Ministry of Education, Korea (Brain Korea 21 Four, Future Veterinary Medicine Leading Education and Research Center), Research Institute for Veterinary Science, College of Veterinary Medicine, Seoul National University, and the Japan Society for the Promotion of Science (JSPS) KAKENHI, Japan (grant no. JP21K18165 and JP22H00200 to H.H.).

References

1. Semenza GL. Hypoxia-inducible factors in physiology and medicine. Cell 2012;148:399–408.
2. Carmeliet P. Blood vessels and nerves: common signals, pathways and diseases. Nat Rev Genet 2003;4:710–20.
3. Kaelin WG Jr, Ratcliffe PJ. Oxygen sensing by metazoans: the central role of the HIF hydroxylase pathway. Mol Cell 2008;30:393–402.
4. Hong BJ, Kim J, Jeong H, Bok S, Kim YE, Ahn GO. Tumor hypoxia and reoxygenation: the yin and yang for radiotherapy. Radiat Oncol J 2016;34:239–49.
5. Chen Z, Han F, Du Y, Shi H, Zhou W. Hypoxic microenvironment in cancer: molecular mechanisms and therapeutic interventions. Signal Transduct Target Ther 2023;8:70.
6. Lau J, Lin KS, Benard F. Past, present, and future: development of theranostic agents targeting carbonic anhydrase IX. Theranostics 2017;7:4322–39.
7. Zandberg DP, Menk AV, Velez M, Normolle D, DePeaux K, Liu A, et al. Tumor hypoxia is associated with resistance to PD-1 blockade in squamous cell carcinoma of the head and neck. J Immunother Cancer 2021;9e002088.
8. Estephan H, Tailor A, Parker R, Kreamer M, Papandreou I, Campo L, et al. Hypoxia promotes tumor immune evasion by suppressing MHC-I expression and antigen presentation. EMBO J 2025;44:903–22.
9. Li Y, Patel SP, Roszik J, Qin Y. Hypoxia-driven immunosuppressive metabolites in the tumor microenvironment: new approaches for combinational immunotherapy. Front Immunol 2018;9:1591.
10. Casazza A, Laoui D, Wenes M, Rizzolio S, Bassani N, Mambretti M, et al. Impeding macrophage entry into hypoxic tumor areas by Sema3A/Nrp1 signaling blockade inhibits angiogenesis and restores antitumor immunity. Cancer Cell 2013;24:695–709.
11. Barsoum IB, Smallwood CA, Siemens DR, Graham CH. A mechanism of hypoxia-mediated escape from adaptive immunity in cancer cells. Cancer Res 2014;74:665–74.
12. Noman MZ, Desantis G, Janji B, Hasmim M, Karray S, Dessen P, et al. PD-L1 is a novel direct target of HIF-1alpha, and its blockade under hypoxia enhanced MDSC-mediated T cell activation. J Exp Med 2014;211:781–90.
13. Palsson-McDermott EM, Dyck L, Zaslona Z, Menon D, McGettrick AF, Mills KH, et al. Pyruvate kinase M2 is required for the expression of the immune checkpoint PD-L1 in immune cells and tumors. Front Immunol 2017;8:1300.
14. Topalian SL, Drake CG, Pardoll DM. Immune checkpoint blockade: a common denominator approach to cancer therapy. Cancer Cell 2015;27:450–61.
15. Wang GL, Semenza GL. Purification and characterization of hypoxia-inducible factor 1. J Biol Chem 1995;270:1230–7.
16. Wang GL, Jiang BH, Rue EA, Semenza GL. Hypoxia-inducible factor 1 is a basic-helix-loop-helix-PAS heterodimer regulated by cellular O2 tension. Proc Natl Acad Sci U S A 1995;92:5510–4.
17. Duan C. Hypoxia-inducible factor 3 biology: complexities and emerging themes. Am J Physiol Cell Physiol 2016;310:C260–9.
18. Jaakkola P, Mole DR, Tian YM, Wilson MI, Gielbert J, Gaskell SJ, et al. Targeting of HIF-alpha to the von Hippel-Lindau ubiquitylation complex by O2-regulated prolyl hydroxylation. Science 2001;292:468–72.
19. Taylor CT, Scholz CC. The effect of HIF on metabolism and immunity. Nat Rev Nephrol 2022;18:573–87.
20. Ivan M, Kondo K, Yang H, Kim W, Valiando J, Ohh M, et al. HIFalpha targeted for VHL-mediated destruction by proline hydroxylation: implications for O2 sensing. Science 2001;292:464–8.
21. Frede S, Stockmann C, Freitag P, Fandrey J. Bacterial lipopolysaccharide induces HIF-1 activation in human monocytes via p44/42 MAPK and NF-kappaB. Biochem J 2006;396:517–27.
22. Potteti HR, Noone PM, Tamatam CR, Ankireddy A, Noel S, Rabb H, et al. Nrf2 mediates hypoxia-inducible HIF1alpha activation in kidney tubular epithelial cells. Am J Physiol Renal Physiol 2021;320:F464–74.
23. Maxwell PH, Wiesener MS, Chang GW, Clifford SC, Vaux EC, Cockman ME, et al. The tumour suppressor protein VHL targets hypoxia-inducible factors for oxygen-dependent proteolysis. Nature 1999;399:271–5.
24. Patel SA, Simon MC. Biology of hypoxia-inducible factor-2alpha in development and disease. Cell Death Differ 2008;15:628–34.
25. Jaskiewicz M, Moszynska A, Kroliczewski J, Cabaj A, Bartoszewska S, Charzynska A, et al. The transition from HIF-1 to HIF-2 during prolonged hypoxia results from reactivation of PHDs and HIF1A mRNA instability. Cell Mol Biol Lett 2022;27:109.
26. Chen N, Hao C, Liu BC, Lin H, Wang C, Xing C, et al. Roxadustat treatment for anemia in patients undergoing long-term dialysis. N Engl J Med 2019;381:1011–22.
27. Haase VH. Regulation of erythropoiesis by hypoxia-inducible factors. Blood Rev 2013;27:41–53.
28. Kaelin WG Jr. The VHL tumor suppressor gene: insights into oxygen sensing and cancer. Trans Am Clin Climatol Assoc 2017;128:298–307.
29. Heikkila M, Pasanen A, Kivirikko KI, Myllyharju J. Roles of the human hypoxia-inducible factor (HIF)-3alpha variants in the hypoxia response. Cell Mol Life Sci 2011;68:3885–901.
30. Maynard MA, Evans AJ, Hosomi T, Hara S, Jewett MA, Ohh M. Human HIF-3alpha4 is a dominant-negative regulator of HIF-1 and is down-regulated in renal cell carcinoma. FASEB J 2005;19:1396–406.
31. Zhang P, Yao Q, Lu L, Li Y, Chen PJ, Duan C. Hypoxia-inducible factor 3 is an oxygen-dependent transcription activator and regulates a distinct transcriptional response to hypoxia. Cell Rep 2014;6:1110–21.
32. Yang SL, Wu C, Xiong ZF, Fang X. Progress on hypoxia-inducible factor-3: Its structure, gene regulation and biological function (Review). Mol Med Rep 2015;12:2411–6.
33. Caldwell CC, Kojima H, Lukashev D, Armstrong J, Farber M, Apasov SG, et al. Differential effects of physiologically relevant hypoxic conditions on T lymphocyte development and effector functions. J Immunol 2001;167:6140–9.
34. Sitkovsky M, Lukashev D. Regulation of immune cells by local-tissue oxygen tension: HIF1 alpha and adenosine receptors. Nat Rev Immunol 2005;5:712–21.
35. Semenza GL. Oxygen sensing, homeostasis, and disease. N Engl J Med 2011;365:537–47.
36. Staples KJ, Sotoodehnejadnematalahi F, Pearson H, Frankenberger M, Francescut L, Ziegler-Heitbrock L, et al. Monocyte-derived macrophages matured under prolonged hypoxia transcriptionally up-regulate HIF-1alpha mRNA. Immunobiology 2011;216:832–9.
37. Raskov H, Orhan A, Christensen JP, Gogenur I. Cytotoxic CD8(+) T cells in cancer and cancer immunotherapy. Br J Cancer 2021;124:359–67.
38. Scharping NE, Rivadeneira DB, Menk AV, Vignali PDA, Ford BR, Rittenhouse NL, et al. Mitochondrial stress induced by continuous stimulation under hypoxia rapidly drives T cell exhaustion. Nat Immunol 2021;22:205–15.
39. Varia MA, Calkins-Adams DP, Rinker LH, Kennedy AS, Novotny DB, Fowler WC Jr, et al. Pimonidazole: a novel hypoxia marker for complementary study of tumor hypoxia and cell proliferation in cervical carcinoma. Gynecol Oncol 1998;71:270–7.
40. Saragovi A, Abramovich I, Omar I, Arbib E, Toker O, Gottlieb E, et al. Systemic hypoxia inhibits T cell response by limiting mitobiogenesis via matrix substrate-level phosphorylation arrest. Elife 2020;9e56612.
41. Vignali PDA, DePeaux K, Watson MJ, Ye C, Ford BR, Lontos K, et al. Hypoxia drives CD39-dependent suppressor function in exhausted T cells to limit antitumor immunity. Nat Immunol 2023;24:267–79.
42. Antonioli L, Pacher P, Vizi ES, Hasko G. CD39 and CD73 in immunity and inflammation. Trends Mol Med 2013;19:355–67.
43. Cunha PP, Minogue E, Krause LC, Hess RM, Bargiela D, Wadsworth BJ, et al. Oxygen levels at the time of activation determine T cell persistence and immunotherapeutic efficacy. Elife 2023;12e84280.
44. Xu Y, Chaudhury A, Zhang M, Savoldo B, Metelitsa LS, Rodgers J, et al. Glycolysis determines dichotomous regulation of T cell subsets in hypoxia. J Clin Invest 2016;126:2678–88.
45. Moser M, Murphy KM. Dendritic cell regulation of TH1-TH2 development. Nat Immunol 2000;1:199–205.
46. Krishnamoorthy N, Oriss T, Paglia M, Ray A, Ray P. A critical role for IL-6 secretion by dendritic cells promoting Th2 and limiting Th1 response (95.24). J Immunol 2007;178(1 Suppl):S181.
47. Ricciardi A, Elia AR, Cappello P, Puppo M, Vanni C, Fardin P, et al. Transcriptome of hypoxic immature dendritic cells: modulation of chemokine/receptor expression. Mol Cancer Res 2008;6:175–85.
48. Elia AR, Cappello P, Puppo M, Fraone T, Vanni C, Eva A, et al. Human dendritic cells differentiated in hypoxia down-modulate antigen uptake and change their chemokine expression profile. J Leukoc Biol 2008;84:1472–82.
49. Kohler T, Reizis B, Johnson RS, Weighardt H, Forster I. Influence of hypoxia-inducible factor 1alpha on dendritic cell differentiation and migration. Eur J Immunol 2012;42:1226–36.
50. Bosco MC, Pierobon D, Blengio F, Raggi F, Vanni C, Gattorno M, et al. Hypoxia modulates the gene expression profile of immunoregulatory receptors in human mature dendritic cells: identification of TREM-1 as a novel hypoxic marker in vitro and in vivo. Blood 2011;117:2625–39.
51. Klesney-Tait J, Turnbull IR, Colonna M. The TREM receptor family and signal integration. Nat Immunol 2006;7:1266–73.
52. Rama I, Bruene B, Torras J, Koehl R, Cruzado JM, Bestard O, et al. Hypoxia stimulus: an adaptive immune response during dendritic cell maturation. Kidney Int 2008;73:816–25.
53. Weigert A, Weichand B, Sekar D, Sha W, Hahn C, Mora J, et al. HIF-1alpha is a negative regulator of plasmacytoid DC development in vitro and in vivo. Blood 2012;120:3001–6.
54. Hacker C, Kirsch RD, Ju XS, Hieronymus T, Gust TC, Kuhl C, et al. Transcriptional profiling identifies Id2 function in dendritic cell development. Nat Immunol 2003;4:380–6.
55. Chambers AM, Matosevic S. Immunometabolic dysfunction of natural killer cells mediated by the hypoxia-CD73 axis in solid tumors. Front Mol Biosci 2019;6:60.
56. Sarkar S, Germeraad WT, Rouschop KM, Steeghs EM, van Gelder M, Bos GM, et al. Hypoxia induced impairment of NK cell cytotoxicity against multiple myeloma can be overcome by IL-2 activation of the NK cells. PLoS One 2013;8e64835.
57. Solocinski K, Padget MR, Fabian KP, Wolfson B, Cecchi F, Hembrough T, et al. Overcoming hypoxia-induced functional suppression of NK cells. J Immunother Cancer 2020;8e000246.
58. Kennedy PR, Arvindam US, Phung SK, Ettestad B, Feng X, Li Y, et al. Metabolic programs drive function of therapeutic NK cells in hypoxic tumor environments. Sci Adv 2024;10eadn1849.
59. Dong H, Strome SE, Salomao DR, Tamura H, Hirano F, Flies DB, et al. Tumor-associated B7-H1 promotes T-cell apoptosis: a potential mechanism of immune evasion. Nat Med 2002;8:793–800.
60. Chen S, Crabill GA, Pritchard TS, McMiller TL, Wei P, Pardoll DM, et al. Mechanisms regulating PD-L1 expression on tumor and immune cells. J Immunother Cancer 2019;7:305.
61. Oh SA, Wu DC, Cheung J, Navarro A, Xiong H, Cubas R, et al. PD-L1 expression by dendritic cells is a key regulator of T-cell immunity in cancer. Nat Cancer 2020;1:681–91.
62. Wang L, Guo W, Guo Z, Yu J, Tan J, Simons DL, et al. PD-L1-expressing tumor-associated macrophages are immunostimulatory and associate with good clinical outcome in human breast cancer. Cell Rep Med 2024;5:101420.
63. Gordon SR, Maute RL, Dulken BW, Hutter G, George BM, McCracken MN, et al. PD-1 expression by tumour-associated macrophages inhibits phagocytosis and tumour immunity. Nature 2017;545:495–9.
64. Hussain K, Liu R, Smith RC, Muller KT, Ghorbani M, Macari S, et al. HIF activation enhances FcgammaRIIb expression on mononuclear phagocytes impeding tumor targeting antibody immunotherapy. J Exp Clin Cancer Res 2022;41:131.
65. Arlauckas SP, Garris CS, Kohler RH, Kitaoka M, Cuccarese MF, Yang KS, et al. In vivo imaging reveals a tumor-associated macrophage-mediated resistance pathway in anti-PD-1 therapy. Sci Transl Med 2017;9eaal3604.
66. Xun Z, Zhou H, Shen M, Liu Y, Sun C, Du Y, et al. Identification of hypoxia-ALCAM(high) macrophage-exhausted T cell axis in tumor microenvironment remodeling for immunotherapy resistance. Adv Sci (Weinh) 2024;11e2309885.
67. Brooks JM, Menezes AN, Ibrahim M, Archer L, Lal N, Bagnall CJ, et al. Development and validation of a combined hypoxia and immune prognostic classifier for head and neck cancer. Clin Cancer Res 2019;25:5315–28.
68. Mayer A, Haist M, Loquai C, Grabbe S, Rapp M, Roth W, et al. Role of hypoxia and the adenosine system in immune evasion and prognosis of patients with brain metastases of melanoma: a multiplex whole slide immunofluorescence study. Cancers (Basel) 2020;12:3753.
69. Xiao X, Guo Q, Cui C, Lin Y, Zhang L, Ding X, et al. Multiplexed imaging mass cytometry reveals distinct tumor-immune microenvironments linked to immunotherapy responses in melanoma. Commun Med (Lond) 2022;2:131.
70. Vos JL, Elbers JB, Krijgsman O, Traets JJ, Qiao X, van der Leun AM, et al. Neoadjuvant immunotherapy with nivolumab and ipilimumab induces major pathological responses in patients with head and neck squamous cell carcinoma. Nat Commun 2021;12:7348.
71. Sanduleanu S, Wiel A, Lieverse RI, Marcus D, Ibrahim A, Primakov S, et al. Hypoxia PET imaging with [18F]-HX4-A promising next-generation tracer. Cancers (Basel) 2020;12:1322.
72. Zheng Y, Yang Y, Xiong Q, Ma Y, Zhu Q. Establishment and verification of a novel gene signature connecting hypoxia and lactylation for predicting prognosis and immunotherapy of pancreatic ductal adenocarcinoma patients by integrating multi-machine learning and single-cell analysis. Int J Mol Sci 2024;25:11143.
73. Wang Y, Wang X, Liu Y, Xu J, Zhu J, Zheng Y, et al. A novel hypoxia- and lactate metabolism-related prognostic signature to characterize the immune landscape and predict immunotherapy response in osteosarcoma. Front Immunol 2024;15:1467052.
74. Li J, Qiao H, Wu F, Sun S, Feng C, Li C, et al. A novel hypoxia-and lactate metabolism-related signature to predict prognosis and immunotherapy responses for breast cancer by integrating machine learning and bioinformatic analyses. Front Immunol 2022;13:998140.
75. Deng C, Deng G, Chu H, Chen S, Chen X, Li X, et al. Construction of a hypoxia-immune-related prognostic panel based on integrated single-cell and bulk RNA sequencing analyses in gastric cancer. Front Immunol 2023;14:1140328.
76. Bai S, Chen L, Yan Y, Wang X, Jiang A, Li R, et al. Identification of hypoxia-immune-related gene signatures and construction of a prognostic model in kidney renal clear cell carcinoma. Front Cell Dev Biol 2021;9:796156.
77. Liu L, Han L, Dong L, He Z, Gao K, Chen X, et al. The hypoxia-associated genes in immune infiltration and treatment options of lung adenocarcinoma. PeerJ 2023;11e15621.
78. Li SR, Man QW, Liu B. Development and validation of a novel hypoxia-related signature for prognostic and immunogenic evaluation in head and neck squamous cell carcinoma. Front Oncol 2022;12:943945.
79. Nie C, Qin H, Zhang L. Identification and validation of a prognostic signature related to hypoxic tumor microenvironment in cervical cancer. PLoS One 2022;17e0269462.
80. Abdelfattah N, Kumar P, Wang C, Leu JS, Flynn WF, Gao R, et al. Single-cell analysis of human glioma and immune cells identifies S100A4 as an immunotherapy target. Nat Commun 2022;13:767.
81. Shou Y, Yang L, Yang Y, Zhu X, Li F, Xu J. Determination of hypoxia signature to predict prognosis and the tumor immune microenvironment in melanoma. Mol Omics 2021;17:307–16.
82. Zhang L, Wang S, Wang Y, Zhao W, Zhang Y, Zhang N, et al. Effects of hypoxia in intestinal tumors on immune cell behavior in the tumor microenvironment. Front Immunol 2021;12:645320.
83. Hu G, Niu W, Ge J, Xuan J, Liu Y, Li M, et al. Identification of thyroid cancer biomarkers using WGCNA and machine learning. Eur J Med Res 2025;30:244.
84. Li K, Yang Y, Ma M, Lu S, Li J. Hypoxia-based classification and prognostic signature for clinical management of hepatocellular carcinoma. World J Surg Oncol 2023;21:216.
85. Chen H, Zhang Y, Chen X, Xu R, Zhu Y, He D, et al. Hypoxia is correlated with the tumor immune microenvironment: potential application of immunotherapy in bladder cancer. Cancer Med 2023;12:22333–53.
86. Cao M, Xiao L, Chen S, Huang J. Characterization of hypoxia-responsive states in ovarian cancer to identify hot tumors and aid adjuvant therapy. Discov Oncol 2024;15:23.
87. Zannella VE, Dal Pra A, Muaddi H, McKee TD, Stapleton S, Sykes J, et al. Reprogramming metabolism with metformin improves tumor oxygenation and radiotherapy response. Clin Cancer Res 2013;19:6741–50.
88. Scharping NE, Menk AV, Whetstone RD, Zeng X, Delgoffe GM. Efficacy of PD-1 blockade is potentiated by metformin-induced reduction of tumor hypoxia. Cancer Immunol Res 2017;5:9–16.
89. Hatfield SM, Kjaergaard J, Lukashev D, Schreiber TH, Belikoff B, Abbott R, et al. Immunological mechanisms of the antitumor effects of supplemental oxygenation. Sci Transl Med 2015;7:277ra30.
90. Chauhan VP, Martin JD, Liu H, Lacorre DA, Jain SR, Kozin SV, et al. Angiotensin inhibition enhances drug delivery and potentiates chemotherapy by decompressing tumour blood vessels. Nat Commun 2013;4:2516.
91. Datta M, Chatterjee S, Perez EM, Gritsch S, Roberge S, Duquette M, et al. Losartan controls immune checkpoint blocker-induced edema and improves survival in glioblastoma mouse models. Proc Natl Acad Sci U S A 2023;120e2219199120.
92. Kakuta H, Sudoh K, Sasamata M, Yamagishi S. Telmisartan has the strongest binding affinity to angiotensin II type 1 receptor: comparison with other angiotensin II type 1 receptor blockers. Int J Clin Pharmacol Res 2005;25:41–6.
93. Wadsworth BJ, Cederberg RA, Lee CM, Firmino NS, Franks SE, Pan J, et al. Angiotensin II type 1 receptor blocker telmisartan inhibits the development of transient hypoxia and improves tumour response to radiation. Cancer Lett 2020;493:31–40.
94. Hunter FW, Young RJ, Shalev Z, Vellanki RN, Wang J, Gu Y, et al. Identification of P450 oxidoreductase as a major determinant of sensitivity to hypoxia-activated prodrugs. Cancer Res 2015;75:4211–23.
95. Wilson WR, Hay MP. Targeting hypoxia in cancer therapy. Nat Rev Cancer 2011;11:393–410.
96. Wang M, He M, Zhang M, Xue S, Xu T, Zhao Y, et al. Controllable hypoxia-activated chemotherapy as a dual enhancer for synergistic cancer photodynamic immunotherapy. Biomaterials 2023;301:122257.
97. Fu Z, Mowday AM, Smaill JB, Hermans IF, Patterson AV. Tumour hypoxia-mediated immunosuppression: mechanisms and therapeutic approaches to improve cancer immunotherapy. Cells 2021;10:1006.
98. Van Cutsem E, Lenz HJ, Furuse J, Tabernero J, Heinemann V, Ioka T, et al. MAESTRO: a randomized, double-blind phase III study of evofosfamide (Evo) in combination with gemcitabine (Gem) in previously untreated patients (pts) with metastatic or locally advanced unresectable pancreatic ductal adenocarcinoma (PDAC). J Clin Oncol 2016;34(15 Suppl):4007.
99. Tap WD, Papai Z, Van Tine BA, Attia S, Ganjoo KN, Jones RL, et al. Doxorubicin plus evofosfamide versus doxorubicin alone in locally advanced, unresectable or metastatic soft-tissue sarcoma (TH CR-406/SARC021): an international, multicentre, open-label, randomised phase 3 trial. Lancet Oncol 2017;18:1089–103.
100. Jayaprakash P, Ai M, Liu A, Budhani P, Bartkowiak T, Sheng J, et al. Targeted hypoxia reduction restores T cell infiltration and sensitizes prostate cancer to immunotherapy. J Clin Invest 2018;128:5137–49.
101. Salman S, Meyers DJ, Wicks EE, Lee SN, Datan E, Thomas AM, et al. HIF inhibitor 32-134D eradicates murine hepatocellular carcinoma in combination with anti-PD1 therapy. J Clin Invest 2022;132e156774.
102. Koh MY, Spivak-Kroizman T, Venturini S, Welsh S, Williams RR, Kirkpatrick DL, et al. Molecular mechanisms for the activity of PX-478, an antitumor inhibitor of the hypoxia-inducible factor-1alpha. Mol Cancer Ther 2008;7:90–100.
103. Luo F, Lu FT, Cao JX, Ma WJ, Xia ZF, Zhan JH, et al. HIF-1alpha inhibition promotes the efficacy of immune checkpoint blockade in the treatment of non-small cell lung cancer. Cancer Lett 2022;531:39–56.
104. Semenza GL. Evaluation of HIF-1 inhibitors as anticancer agents. Drug Discov Today 2007;12:853–9.
105. Seifert AM, Kim TS, Greer JB, Cohen NA, Beckman MJ, Santamaria-Barria JA, et al. PD-1/PD-L1 blockade enhances the efficacy of imatinib in gastrointestinal stromal tumor (GIST). J Am Coll Surg 2014;219(3 Suppl):S129.
106. Yang JC, Gadgeel SM, Sequist LV, Wu CL, Papadimitrakopoulou VA, Su WC, et al. Pembrolizumab in combination with erlotinib or gefitinib as first-line therapy for advanced NSCLC with sensitizing EGFR mutation. J Thorac Oncol 2019;14:553–9.
107. Zhou S, Wang B, Wei Y, Dai P, Chen Y, Xiao Y, et al. PD-1 inhibitor combined with Docetaxel exerts synergistic anti-prostate cancer effect in mice by down-regulating the expression of PI3K/AKT/NFKB-P65/PD-L1 signaling pathway. Cancer Biomark 2024;40:47–59.
108. Onnis B, Rapisarda A, Melillo G. Development of HIF-1 inhibitors for cancer therapy. J Cell Mol Med 2009;13:2780–6.
109. Wang T, Lei Y, Sun J, Wang L, Lin Y, Wu Z, et al. Targeting NANOS1 in triple-negative breast cancer: synergistic effects of digoxin and PD-1 inhibitors in modulating the tumor immune microenvironment. Front Oncol 2024;14:1536406.
110. Kosti P, Opzoomer JW, Larios-Martinez KI, Henley-Smith R, Scudamore CL, Okesola M, et al. Hypoxia-sensing CAR T cells provide safety and efficacy in treating solid tumors. Cell Rep Med 2021;2:100227.
111. He H, Liao Q, Zhao C, Zhu C, Feng M, Liu Z, et al. Conditioned CAR-T cells by hypoxia-inducible transcription amplification (HiTA) system significantly enhances systemic safety and retains antitumor efficacy. J Immunother Cancer 2021;9e002755.
112. Zhu X, Chen J, Li W, Xu Y, Shan J, Hong J, et al. Hypoxia-responsive CAR-T cells exhibit reduced exhaustion and enhanced efficacy in solid tumors. Cancer Res 2024;84:84–100.
113. Liao Q, He H, Mao Y, Ding X, Zhang X, Xu J. Engineering T cells with hypoxia-inducible chimeric antigen receptor (HiCAR) for selective tumor killing. Biomark Res 2020;8:56.
114. Hegde A, Jayaprakash P, Couillault CA, Piha-Paul S, Karp D, Rodon J, et al. A phase I dose-escalation study to evaluate the safety and tolerability of evofosfamide in combination with ipilimumab in advanced solid malignancies. Clin Cancer Res 2021;27:3050–60.
115. Hutt DM, Roth DM, Vignaud H, Cullin C, Bouchecareilh M. The histone deacetylase inhibitor, Vorinostat, represses hypoxia inducible factor 1 alpha expression through translational inhibition. PLoS One 2014;9e106224.
116. Zhang C, Yang C, Feldman MJ, Wang H, Pang Y, Maggio DM, et al. Vorinostat suppresses hypoxia signaling by modulating nuclear translocation of hypoxia inducible factor 1 alpha. Oncotarget 2017;8:56110–25.
117. Rodriguez CP, Wu QV, Voutsinas J, Fromm JR, Jiang X, Pillarisetty VG, et al. A phase II trial of pembrolizumab and vorinostat in recurrent metastatic head and neck squamous cell carcinomas and salivary gland cancer. Clin Cancer Res 2020;26:837–45.
118. Yoshihara T, Hosaka M, Terata M, Ichikawa K, Murayama S, Tanaka A, et al. Intracellular and in vivo oxygen sensing using phosphorescent Ir(III) complexes with a modified acetylacetonato ligand. Anal Chem 2015;87:2710–7.
119. Parpaleix A, Goulam Houssen Y, Charpak S. Imaging local neuronal activity by monitoring PO(2) transients in capillaries. Nat Med 2013;19:241–6.
120. Beinlich FR, Asiminas A, Untiet V, Bojarowska Z, Pla V, Sigurdsson B, et al. Oxygen imaging of hypoxic pockets in the mouse cerebral cortex. Science 2024;383:1471–8.
121. O’Connor JP, Robinson SP, Waterton JC. Imaging tumour hypoxia with oxygen-enhanced MRI and BOLD MRI. Br J Radiol 2019;92:20180642.
122. Logothetis NK. What we can do and what we cannot do with fMRI. Nature 2008;453:869–78.
123. Hallac RR, Ding Y, Yuan Q, McColl RW, Lea J, Sims RD, et al. Oxygenation in cervical cancer and normal uterine cervix assessed using blood oxygenation level-dependent (BOLD) MRI at 3T. NMR Biomed 2012;25:1321–30.
124. Elas M, Williams BB, Parasca A, Mailer C, Pelizzari CA, Lewis MA, et al. Quantitative tumor oxymetric images from 4D electron paramagnetic resonance imaging (EPRI): methodology and comparison with blood oxygen level-dependent (BOLD) MRI. Magn Reson Med 2003;49:682–91.
125. Padhani AR. Dynamic contrast-enhanced MRI in clinical oncology: current status and future directions. J Magn Reson Imaging 2002;16:407–22.
126. Tao X, Wang L, Hui Z, Liu L, Ye F, Song Y, et al. DCE-MRI perfusion and permeability parameters as predictors of tumor response to CCRT in patients with locally advanced NSCLC. Sci Rep 2016;6:35569.
127. Fowler AM, Strigel RM. Clinical advances in PET-MRI for breast cancer. Lancet Oncol 2022;23:e32–43.
128. Hendrickson K, Phillips M, Smith W, Peterson L, Krohn K, Rajendran J. Hypoxia imaging with [F-18] FMISO-PET in head and neck cancer: potential for guiding intensity modulated radiation therapy in overcoming hypoxia-induced treatment resistance. Radiother Oncol 2011;101:369–75.
129. Wang F, Wan H, Ma Z, Zhong Y, Sun Q, Tian Y, et al. Lightsheet microscopy in the near-infrared II window. Nat Methods 2019;16:545–52.
130. Cochran JM, Busch DR, Leproux A, Zhang Z, O’Sullivan TD, Cerussi AE, et al. Tissue oxygen saturation predicts response to breast cancer neoadjuvant chemotherapy within 10 days of treatment. J Biomed Opt 2018;24:1–11.
131. Poncelet M, Huffman JL, Khramtsov VV, Dhimitruka I, Driesschaert B. Synthesis of hydroxyethyl tetrathiatriarylmethyl radicals OX063 and OX071. RSC Adv 2019;9:35073–6.
132. Epel B, Bowman MK, Mailer C, Halpern HJ. Absolute oxygen R1e imaging in vivo with pulse electron paramagnetic resonance. Magn Reson Med 2014;72:362–8.
133. Wojtkowiak JW, Cornnell HC, Matsumoto S, Saito K, Takakusagi Y, Dutta P, et al. Pyruvate sensitizes pancreatic tumors to hypoxia-activated prodrug TH-302. Cancer Metab 2015;3:2.
134. Gertsenshteyn I, Epel B, Ahluwalia A, Kim H, Fan X, Barth E, et al. The optimal (18)F-fluoromisonidazole PET threshold to define tumor hypoxia in preclinical squamous cell carcinomas using pO(2) electron paramagnetic resonance imaging as reference truth. Eur J Nucl Med Mol Imaging 2022;49:4014–24.
135. Shao Q, Ashkenazi S. Photoacoustic lifetime imaging for direct in vivo tissue oxygen monitoring. J Biomed Opt 2015;20:036004.
136. Wang Y, Hu S, Maslov K, Zhang Y, Xia Y, Wang LV. In vivo integrated photoacoustic and confocal microscopy of hemoglobin oxygen saturation and oxygen partial pressure. Opt Lett 2011;36:1029–31.
137. Palazon A, Goldrath AW, Nizet V, Johnson RS. HIF transcription factors, inflammation, and immunity. Immunity 2014;41:518–28.
138. McNeal KC, Reeves KM, Song PN, Lapi SE, Sorace AG, Larimer BM. [(18)F]FMISO-PET imaging reveals the role of hypoxia severity in checkpoint blockade response. Nucl Med Biol 2024;134-135:108918.
139. Constantinidou A, van der Graaf WT. The fate of new fosfamides in phase III studies in advanced soft tissue sarcoma. Eur J Cancer 2017;84:257–61.
140. Reeves KM, Song PN, Angermeier A, Della Manna D, Li Y, Wang J, et al. (18)F-FMISO PET imaging identifies hypoxia and immunosuppressive tumor microenvironments and guides targeted evofosfamide therapy in tumors refractory to PD-1 and CTLA-4 inhibition. Clin Cancer Res 2022;28:327–37.
141. Pian Q, Li B, Sencan-Egilmez I, Cheng X, Dubb J, Huang X, et al. Out-of-focus signal rejection for in vivo pO(2) measurements using two-photon phosphorescence lifetime microscopy. Biomed Opt Express 2025;16:159–76.
142. Kim Y, Park JJ, Kim CK. Blood oxygenation level-dependent MRI at 3T for differentiating prostate cancer from benign tissue: a preliminary experience. Br J Radiol 2022;95:20210461.
143. Nguyen T, Park S, Hill B, Gandjbakhche AH. Single source-detector separation approach to calculate tissue oxygen saturation using continuous wave near-infrared spectroscopy. IEEE Open J Eng Med Biol 2023;4:79–84.
144. Lee CW, Stantz KM. Development of a mathematical model to estimate intra-tumor oxygen concentrations through multi-parametric imaging. Biomed Eng Online 2016;15:114.
145. Virani N, Kwon J, Zhou H, Mason R, Berbeco R, Protti A. In vivo hypoxia characterization using blood oxygen level dependent magnetic resonance imaging in a preclinical glioblastoma mouse model. Magn Reson Imaging 2021;76:52–60.
146. Vitorino R, Barros AS, Guedes S, Caixeta DC, Sabino-Silva R. Diagnostic and monitoring applications using near infrared (NIR) spectroscopy in cancer and other diseases. Photodiagnosis Photodyn Ther 2023;42:103633.
147. Schaner PE, Pettus JR, Flood AB, Williams BB, Jarvis LA, Chen EY, et al. OxyChip implantation and subsequent electron paramagnetic resonance oximetry in human tumors is safe and feasible: first experience in 24 patients. Front Oncol 2020;10:572060.
148. Steinberg I, Huland DM, Vermesh O, Frostig HE, Tummers WS, Gambhir SS. Photoacoustic clinical imaging. Photoacoustics 2019;14:77–98.
149. Canavesi I, Viswakarma N, Epel B, McMillan A, Kotecha M. Accelerated EPR imaging using deep learning denoising. Magn Reson Med 2025;94:436–46.
150. Jardim-Perassi BV, Mu W, Huang S, Tomaszewski MR, Poleszczuk J, Abdalah MA, et al. Deep-learning and MR images to target hypoxic habitats with evofosfamide in preclinical models of sarcoma. Theranostics 2021;11:5313–29.
151. Matsumoto S, Kishimoto S, Saito K, Takakusagi Y, Munasinghe JP, Devasahayam N, et al. Metabolic and physiologic imaging biomarkers of the tumor microenvironment predict treatment outcome with radiation or a hypoxia-activated prodrug in mice. Cancer Res 2018;78:3783–92.
152. Epel B, Maggio MC, Barth ED, Miller RC, Pelizzari CA, Krzykawska-Serda M, et al. Oxygen-guided radiation therapy. Int J Radiat Oncol Biol Phys 2019;103:977–84.
153. Jo J, Folz J, Gonzalez ME, Paoli A, Eido A, Salfi E, et al. Personalized oncology by in vivo chemical imaging: photoacoustic mapping of tumor oxygen predicts radiotherapy efficacy. ACS Nano 2023;17:4396–403.
154. Wu B, Zhang B, Li B, Wu H, Jiang M. Cold and hot tumors: from molecular mechanisms to targeted therapy. Signal Transduct Target Ther 2024;9:274.
155. Lu Y, Houson HA, Gallegos CA, Mascioni A, Jia F, Aivazian A, et al. Evaluating the immunologically “cold” tumor microenvironment after treatment with immune checkpoint inhibitors utilizing PET imaging of CD4 + and CD8 + T cells in breast cancer mouse models. Breast Cancer Res 2024;26:104.
156. Dewhirst MW, Mowery YM, Mitchell JB, Cherukuri MK, Secomb TW. Rationale for hypoxia assessment and amelioration for precision therapy and immunotherapy studies. J Clin Invest 2019;129:489–91.
157. Chitneni SK, Palmer GM, Zalutsky MR, Dewhirst MW. Molecular imaging of hypoxia. J Nucl Med 2011;52:165–8.

Article information Continued

Copyright © 2026 by the Korean Cancer Association

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Fig. 1.

Fig. 1.

A schematic diagram demonstrating how tumor hypoxia impacts the functions of CD8+ T cells, dendritic cells, and natural killer cells. CCL, CC chemokine ligand; CCR, CC chemokine receptor; IFN, interferon; MICA/B, MHC class I chain-related molecule A and B; ROS, reactive oxygen species; ULBP-1/6, UL16 binding protein 1-6.

Fig. 2.

Fig. 2.

A diagram demonstrating how various approaches inhibiting tumor hypoxia could improve immunotherapy efficacy. CAR T cells, chimeric antigen receptor T cells; CTLA-4, cytotoxic T-lymphocyte antigen 4; HAP, hypoxia-activated prodrugs; HIF, hypoxia-inducible factor; HRE, hypoxia response elements; NK, natural killer; ODD, oxygen-dependent degradation domain; PD-L1, programmed death-ligand 1; PD-1, programmed death 1.

Fig. 3.

Fig. 3.

In vivo pO2 mapping of mouse xenograft models of pancreatic ductal adenocarcinoma cell lines (A, Hs766t; B, MIA PaCa-2; C, SU.86.86) and anatomic maps with magnetic resonance imaging. Also, pO2 maps after intravenous injection of pyruvate are given at the time points until 60 minutes. Reprinted from Wojtkowiak et al. (2015), Pyruvate sensitizes pancreatic tumors to hypoxia-activated prodrug TH-302. Cancer Metab., volume 3, article no. 2, licensed under Creative Commons Attribution (CC BY 4.0) license [133].

Table 1.

Studies analyzing hypoxia-relevant genes as a biomarker predicting the tumor response to immunotherapy in the patients with various types of cancers

Cancer type Database Method(s) to analyze the relationship between hypoxia and immune cells Hypoxia-relevant gene Result Reference
Osteosarcoma Transcriptional expression data: TARGET, GEO CIBERSORT, Immune cell abundance identifier (ImmuCellAI), ESTIMATE, LASSO regression MAFF, COL5A2, FAM162A, SQOR, UQCRB, SFXN4, PFKFB2, COX6A2 Patients with high risk scores (high hypoxic gene expression) exhibited poor prognosis, ‘cold’ tumor phenotype, predicted to have poor responses to immunotherapy [73]
Transcriptome data: GTEx
Hypoxia-related genes: Molecular Signatures Database (MSigDB)
Breast cancer RNA-seq: TCGA, the University of California Santa Cruz Xena ImmuCellAI, ESTIMATE, Immunophenoscore (IPS), TIDE algorithm DARS2, ESRP1, SLC2A1, TH, MAFF Higher risk group with poor outcomes had lower stromal and immune score with increased M2 & M0 macrophages, resting NK cells, regulatory T cells, monocytes and neutrophils. Prognosis analysis towards anti–PD-L1 therapy did not meet the statistical differences between low and high risk score patient groups [74]
Microarray dataset: GEO
Immunotherapeutic cohorts: IMvigor210 cohort (advanced urothelial cancer with atezolizumab intervention) & GEO
Hypoxia-related genes: MSigDB
Gastric cancer Single-cell RNA-seq: GEO database (10 normal tissue, 26 gastric cancer samples) CIBERSORT, ESTIMATE, LASSO regression, TIDE POSTN, BMP4, MXRA5, LBH Higher hypoxia score patients had fewer antitumor immune cells (e.g., activated NK cells and CD8+ T cells) while cancer-promoting immune cells (e.g., resting NK cells and M2 macrophages) were increased. Patients with high hypoxia score had a poor prognosis to immunotherapy and lower TMB [75]
Bulk RNA-seq: TCGA (32 normal tissues, 375 gastric cancer tissues)
Pancreatic ductal adenocarcinoma RNA-seq: TCGA LASSO regression, XGBoost, Random Forest, TIDE PKP2, SLC2A1, LDHA, PlGA, PLOD1, PTGS2, GCK, POMT1, TIMM50, FUT8, ENO1, PDK1, CHEK2, MPC1, HSPA5, PLAC8, KDM3A, P4HA1, PPlA, NUP98, ALDH1B1, 6TC1, SLC25A4, TIMMDC1, CENPA, NSUN2, SLC25A42, TIPARP, ERRFI1, PNPT1, AK4 Patients with high-risk scores had more severe immunosuppressive environment (e.g., M0 macrophages and dendritic cells) and poor prognosis to immunotherapy [72]
Transcriptome data: GTEx
Clear cell renal carcinoma RNA-seq: TCGA-KIRC sample LASSO regression, t-SNE, ImmuCellAI PLAUR, UCN, SPABPC1L, LC16A12, NFE2L3, KCNAB1 High hypoxia and low immune status are associated with poor overall survival. A prognostic model based on six hypoxia–immune-related genes was constructed and validated, showing good predictive performance [76]
Lung adenocarcinoma RNA-seq: TCGA-LUAD sample LASSO regression HK1, SLC2A1, STC1, XPNPEP1, PDK3, PFKL Hypoxia-related genes that their expression influenced immune cell infiltration can serve as poor prognostic markers and potential targets for immunotherapy in LUAD [77]
Microarray: GEO (GSE72094)
Hypoxia-related genes: MsigDB database (HALLMARK_HYPOXIA)
Head and neck squamous cell carcinoma (HNSCC) RNA-seq: TCGA-HNSCC sample LASSO regression, CIBERSORT, ESTIMATE, TIDE, Xcell SRPX, PGK1, STG1, HS3ST1, CDKN1B, HK1 The group with high expression of hypoxia-related genes showed lower immune cell infiltration and poor prognosis. This prognostic model also has potential for predicting responses to immune checkpoint inhibitor therapy [78]
Microarray: GEO (GSE65858, GSE41613, GSE85446)
Hypoxia-related genes: MsigDB database (HALLMARK_HYPOXIA)
Cervical cancer RNA-seq: TCGA-CESC sample Cox regression, LASSO regression, ESTIMATE, CIBERSORT, MCPcounter EFNA1, IER3, ISG20, KLF7, LDHC, P4HA2, PGM1, RBPJ, STC1 Hypoxia-related genes are associated with increased immunosuppressive cells and decreased antitumor immune cells in tumor. ccHPS model might give us insights into antitumor immunotherapy and further improve the strategies for treating cervical cancer [79]
Microarray: GEO (GSE44001)
Gene annotation: GENCODE (human)
Chemotherapeutic sensitivity for tumor samples: genomics of drug sensitivity in cancer database (GDSC)
Hypoxia-related genes: MsigDB database (HALLMARK_HYPOXIA)
Glioblastoma scRNA-seq: human glioma tissue (18 patients, 44 tumor regions) Single-cell RNA-seq analysis, UMAP, CellPhone DB VEGFA, LDHA, ENO1, CA9, S100A4, P4HA1, ADM Hypoxia-enriched myeloid clusters (MC3-MC5) were associated with immunosuppressive phenotypes and poor survival. Hypoxia influenced immune cell phenotype and spatial distribution in GBM [80]
- Hypoxia-related genes: MSigDB (HALLMARK_HYPOXIA)
Melanoma (skin cutaneous melanoma) RNA-seq: The Cancer Genome Atlas (TCGA-SKCM) CIBERSORT, LASSO regression FBP1, SDC3, FOXO3 IGFBP1, S100A4, EGFR, ISG20, CP, PPARGC1A, KIF5A, DPYSL4 Patients in the high hypoxia score group showed decreased infiltration of CD8+ T cells and activated memory CD4+ T cells, along with increased Tregs and M2 macrophages, indicating an immunosuppressive tumor microenvironment and poor prognosis [81]
- Hypoxia-related genes: MSigDB database
Colon adenocarcinoma RNA-seq: TCGA-COAD CIBERSORT ALDOB, GPC1, ALDOC, SLC2A3 High hypoxia risk scores were associated with increased M0 macrophages, decreased CD8+ T cell infiltration, and poor prognosis. Prognostic signature based on 4 genes was constructed and validated [82]
Microarray: GEO (GSE39582)
Hypoxia gene sets: MSigDB (HALLMARK_HYPOXIA)
Thyroid cancer RNA-seq: TCGA-THCA CIBERSORT, Weighted Gene Co-expression Network Analysis (WGCNA), LASSO regression, machine learning (XGBoost, Random Forest) P4HA2, TFF3, RPS6KA5, EYA1 High P4HA2 expression was correlated with increased M2 macrophages and Tregs, forming an immunosuppressive microenvironment linked to tumor progression [83]
Microarray: GEO (GSE29265, GSE33630)
Hepatocellular carcinoma RNA-seq: TCGA-liver hepatocellular carcinoma CIBERSORT, LASSO regression ANLN, CBX2, DLGAP5, FBLN2, FTCD, HMOX1, IGLV1-44, IL33, LCAT, LPCAT1, MK167, PFN2, RNF145, S100A9, SPP1 Hypoxia subtype cluster 2 (worse overall survival, disease survival, disease-specific specific survival, progression-free survival) displayed increased TIL (tumor-infiltrating lymphocytes) including CD8 T cells and regulatory T cells. [84]
These patients also exhibited higher PD-L1 expression and higher TMB
Bladder cancer RNA-seq: TCGA-BLCA sample ESTIMATE, ssGSEA, MCPcounter, EPIC, TIMER ANXA6, CYBB, SP11, C5AR1, COL6A1, COL6A2, ITGB2, TIMP2, FCER1G, TLR8 Patients with higher hypoxic scores had shorter overall survival while infiltration density and immune score including CD8 T-effector signature were increased [85]
Hypoxia-related genes: GeneCards database
Immunotherapeutic cohorts: IMvigor210 cohort (advanced urothelial cancer with atezolizumab intervention)
Ovarian cancer RNA-seq: TCGA-OV sample ssGSEA, ESTIMATE, Boruta PGAM1, TPI1, SLC2A1, TUBB6, VEGFA, ENO1, ADM, ALDOA, CDKN3, LDHA, MIF, MRPS17, NDRG1, P4HA1 Patients with high hypoxia scores exhibited increased immune cell infiltration and showed greater sensitivity to immunotherapy, displaying characteristics of ‘hot tumors.' Evaluating hypoxia response states may thus aid in developing more effective immunotherapy strategies [86]
Microarray: GEO (GSE18520)
International cancer genome consortium (ICGC-OV-AU),
Transcriptome data: GTEx (normal ovarian tissue)
Hypoxia-related genes: MsigDB database (HALLMARK_HYPOXIA)
Hypoxia score evaluation: CMAP

Note that the rows colored in gray indicate the studies reporting a positive prognosis between hypoxic gene expression and the outcome of patients (i.e., high hypoxic gene signature predicts a better treatment response). BLCA, bladder cancer; CESC, cervical squamous cell carcinoma and endocervical adenocarcinoma; CIBERSORT, Cell type Identification By Estimating Relative Subsets Of RNA Transcripts; CMAP, connectivity map; COAD, colon adenocarcinoma; EPIC, Estimating the Proportion of Immune and Cancer cells; ESTIMATE, Estimation of STromal and Immune cells in MAlignant Tumor tissues using Expression data; GBM, glioblastoma multiforme; GDSC, Genomics of Drug Sensitivity in Cancer; GEO, Gene Expression Omnibus; GTEx, Genotype-Tissue Expression project; HNSCC, head and neck squamous cell carcinoma; ICGC-OV-AU, International Cancer Genome Consortium Ovarian Cancer–Australia project; KIRC, kidney renal clear cell carcinoma; LASSO, Least Absolute Shrinkage and Selection Operator; LUAD, lung adenocarcinoma; OV, ovarian cancer; PD-L1, programmed death-ligand 1; RNA-seq, RNA sequencing; scRNA-seq, single-cell RNA sequencing; SKCM, skin cutaneous melanoma; ssGSEA, single-sample gene set enrichment analysis; TARGET, Therapeutically Applicable Research to Generate Effective Treatments; TCGA, The Cancer Genome Atlas; THCA, thyroid carcinoma; TIDE, Tumor Immune Dysfunction and Exclusion; TIL, tumor-infiltrating lymphocytes; TIMER, Tumor IMmune Estimation Resource; TMB, tumor mutational burden; t-SNE, t-distributed Stochastic Neighbor Embedding; UMAP, Uniform Manifold Approximation and Projection.

Table 2.

Accuracy, sensitivity, specificity, feasibility in the clinical use, and the limitations of non-invasive imaging techniques for tumor hypoxia

Technique Imaging agent requirements Measured entity Measurement accuracy or sensitivity and specificity Feasibility or availability in clinical use Limitations and further comments Reference
Fluorescent/phosphorescent imaging Fluorescent/phosphorescent dyes Hypoxia in living cells, pO2 Phosphorescent lifetime microscopy: pO2 average error 5 mmHg Feasible Limited penetration depth; requirement of the ergulatory agency’s approval for dyes [141]
Blood-oxygen level-dependent magnetic resonance imaging None Difference in magnetization between oxy- and deoxyhemoglobin Sensitivity 73.8%, specificity 52% Available in clinics Not providing tissue pO2 [142]
Dynamic contrast-enhanced MRI Gadolinium-based contrast agent Tissue vasculature and permeability Indirectly associated with hypoxia Available in clinics Not providing tissue pO2 [121]
18F-fluoromisonidazole positron emission tomography 18F-fluoromisonidazole Hypoxia in living cells Sensitivity 100%, specificity 87.5% Available in clinics Radiation exposure; not providing tissue pO2 [140]
Near-infrared spectroscopy and imaging None Optical absorption of oxy- and deoxyhemoglobin, saturation of percutaneous oxygen SpO2, tissue oxygen saturation StO2 StO2 average error 0.3% Feasible Not quantitative for tissue pO2 [143]
A point measurement like a pulse oximeter (SpO2) is routinely used in clinics
Electron paramagnetic resonance (EPR) imaging Free radicals stable in vivo pO2 pO2 resolution 1 mmHg Feasible A human-size EPR imager is not available for clinical use [132]
EPR spectroscopy has been performed in a clinical trial
Requirement of the regulatory agency’s approval for free radical agents
Photoacoustic imaging (PAI) Methylene blue, other PA agents pO2 pO2 average error 7% Feasible Requirement of the regulatory agency’s approval for PA agents [144]

MRI, magnetic resonance imaging; PA, photoacoustic.