Head and neck squamous cell carcinoma (HNSCC) is the most common type of head and neck cancer and causes significant cancer-related mortality and morbidity worldwide [1,2]. Although several treatment approaches can be applied, including surgery, chemotherapy, radiation therapy, and targeted therapy, nearly half of the patients still experience disease relapse or progression after treatment [3]. Treatments targeting immune checkpoints, specifically the programmed death-ligand 1 (PD-L1)/PD-1 axis, improve progression-free survival and overall survival in many solid tumors [4]. Based on satisfactory results from a pivotal trial—KEYNOTE-048—the U.S. Food and Drug Administration (FDA) approved pembrolizumab as the primary therapeutic line for metastatic, recurrent, or unresectable disease in patients with HNSCC. However, the objective response rates (ORRs) for nivolumab and pembrolizumab in HNSCC remain <20 %, which are remarkably low compared with the ORRs of >80 % in lymphoma [5]. This discrepancy highlights the need for additional research for elucidating the mechanisms underlying the PD-1/PD-L1 circuitry in HNSCC, which may lead to improved treatment strategies and enhanced patient outcomes.

T-cell activation and dysfunction are governed by both direct interactions and regulatory mechanisms involving receptors. PD-L1, serving as the primary ligand for PD-1, triggers an inhibitory signal in activated T-cells, leading to apoptosis, anergy (a state of unresponsiveness), and eventual functional exhaustion. The process of inducing T-cell clonal anergy is facilitated by PD-L1, a transmembrane peptide encoded by the CD274 gene. By binding to the inhibitory receptor PD-1, PD-L1's interaction promotes its overexpression in various cancer cells, including those in lung, breast, melanoma, bladder, and head and neck cancers. This overexpression allows these cancer cells to evade immune detection and surveillance [5]. Therapeutic approaches targeting immune checkpoints, including ligand–receptor interactions between PD-L1 and PD-1, have been demonstrated to offer significant clinical benefits for treating human cancers. Nevertheless, many patients with cancer, particularly HNSCC, exhibit inefficient responses to immunotherapies targeting PD-1 or PD-L1 [6]. The specific mechanisms underlying this lack of responsiveness remain unclear and warrant further investigation. A significant link exists between PD-L1 expression in cancer cells and their sensitivity to PD-1–targeting or PD-L1–targeting drugs [7,8]. Consequently, the mechanisms underlying the regulation of PD-L1 expression and maintenance must be elucidated to facilitate the development of more potent immunotherapy strategies for HNSCC.

Ferroptosis, also known as iron-mediated cell death, has gained increased attention in cancer immunotherapy research due to its role in facilitating immunotherapy-mediated cell death, which is independent of other established cell death mechanisms, such as apoptosis, autophagy, and necroptosis [9]. However, cancer cells can suppress ferroptosis by expressing high levels of a cysteine transporter, solute carrier family 7 member 11 (SLC7A11), which is controlled by the nuclear factor erythroid 2-related factor 2 (NRF2), a factor that counteracts the iron-mediated peroxidation of lipid products [10]. NRF2, a major transcription factor, can also suppress ferroptosis by increasing the expression of genes that inhibit ferroptosis, including glutathione peroxidase 4 (GPX4), NAD(P)H dehydrogenase 1 (NQO1), glutathione synthetase (GSS), and glutamate-cysteine ligase catalytic subunit (GCLC); however, the excessive activation of NRF2 may trigger ferroptosis [9,10]. Multiple signaling pathways can be hijacked by ferroptosis to inhibit tumor growth and improve cancer cell immunotherapy. By controlling immune checkpoints, cancer cells can evade immune surveillance by T cells, macrophages, and dendritic cells. Inducing ferroptosis, however, can elevate CD8+ T-cell activity and enhance tumor immunotherapy [11]. Activated CD8+ T cells can inhibit SLC7A11 expression in cancer cells, leading to lipid peroxidation and ferroptosis. Thus, ferroptosis might contribute to T-cell–induced tumor cell death. By contrast, cholesterol- and CD8+ T cell–mediated ferroptosis can enable neoplastic cells to escape immune surveillance. In this process, CD8+ T cells accumulate and encircle tumors, uptake fatty acids, and mediate ferroptosis due to fat-mediated CD36 upregulation [12]. Inhibiting CD36 can enhance anti-PD-1 immunotherapy, which indicates that ferroptosis inducers may impair T-cell survival. The antiporter system Xc− transports cystine and glutamate and is critical for preventing ferroptosis, because it imports cystine to generate reduced glutathione and activate GPX4, which mediates antioxidant defenses, thereby preventing ferroptosis [13]. The specific mechanism through which HNSCC tumors resist immunotherapy-mediated ferroptosis remains unknown, and ferroptosis regulators other than GPX4 may be involved in the mechanism.

Among the several redox proteins found in living organisms is the thioredoxin family comprising thioredoxins (TRXs), glutaredoxins (GLRXs), and peroxiredoxins (PRDXs) [14]. In general, TRXs and GLRXs are disulfide reductases that reduce disulfide bridges in oxidized proteins by using NADPH as the reducing agent; thus, they serve as antioxidant proteins. They contribute to maintaining cellular homeostasis and protein function by reducing PRDXs. Moreover, the TRX system comprises TRX and thioredoxin reductase (TXNRD), which restores the system by reducing oxidized TRX by using electrons from NADPH [15]. TRX is present in the extracellular compartment, and it has various subcellular localizations, including the cytoplasm, mitochondria, and nucleus. Different TRX isoforms are involved in different cellular functions and activate different molecular mechanisms, such as redox balance, cell proliferation, and apoptosis [16]. TRX and TXNRD1 are involved in ferroptosis. TXNRD1 is an indirect indicator of NRF2 pathway activation in certain cancers [17]. Auranofin, a drug initially approved for treating rheumatoid arthritis, is currently under investigation in clinical trials as a promising anti-cancer agent. Its primary mechanism involves targeting the antioxidant system governed by thioredoxin reductase (TrxR), a system that shields cells from oxidative damage and apoptosis in both the cytoplasm and mitochondria. TrxR is found to be upregulated in various cancers, where it supports cancer cell growth as an adaptive strategy, thereby positioning TrxR as a compelling target for cancer therapy and highlighting auranofin's potential as an anti-cancer drug. Auranofin has been shown to inhibit the activity of TXNRD1 by 50 % at a concentration of 0.5 μM [18]. By impeding TrxR, auranofin disrupts the balance of intracellular redox states, leading to an accumulation of reactive oxygen species (ROS) and promoting cell death. Another key action of auranofin involves simulating the inhibition of proteasomes by obstructing the ubiquitin-proteasome system (UPS). The UPS plays a crucial role in several cellular processes vital for cancer cell survival, including cell cycle regulation, protein breakdown, gene expression, and DNA repair. This makes it particularly vital for cancer cells compared to non-cancerous cells. Over the last decade or so, preclinical studies have increasingly supported the anti-cancer potential of auranofin, especially through its ability to elevate oxidative stress within cancer cells. For instance, research has shown that head and neck squamous cell carcinoma cells critically depend on Trx 1 for their survival. These findings were corroborated by experiments where head and neck cancer cells were treated with auranofin, with or without the ROS scavenger N-acetylcysteine (NAC), which mitigates ROS-induced DNA damage. Pretreatment with NAC was found to reverse the lethal effects of auranofin on cancer cells, demonstrating that the modulation of ROS levels is a key mechanism through which auranofin impacts the growth and dissemination of head and neck cancer.

In mice, the TXNRD1 inhibitor auranofin was observed to trigger lipid peroxidation and ferroptosis. Auranofin also increased lysine oxidase (LOX)-mediated cytotoxicity by activating caspase-independent ferroptosis [19]. Although TRXs are believed to play a role in ferroptosis, data on the role of TXNRD1 in cellular resistance to immunotherapy-induced ferroptosis are lacking.

In the present study, we identified a targetable regulator of resistance upon immunotherapy-mediated ferroptosis in HNSCC. We demonstrated that TXNRD1 was aberrantly expressed in HNSCC tumors and closely linked with NRF2 overactivity and PD-L1 expression. We also observed that TXNRD1 binds to ribonucleotide reductase regulatory subunit M2 (RRM2) to mediate PD-L1 transcription in HNSCC cell lines. Disruption of TXNRD1 expression sensitized HNSCC cells and organoids to anti-PD-1-mediated Jurkat T-cell activation, which promoted T-cell tumor killing and cancer cell ferroptosis. Inhibiting TXNRD1 expression through auranofin cotreatment improved the efficacy of PD-1 inhibition by mobilizing CD8+ T cells. Taken together, our findings imply that TXNRD1 inhibition can improve immunotherapy outcomes in patients with HNSCC.