Introduction
Cervical cancer remains one of the most prevalent HPV (human papillomavirus)-driven malignancies worldwide, with persistent high-risk HPV infection serving as the primary etiologic driver. While prophylactic vaccination has reduced incidence in several high-income settings, advanced and recurrent disease continues to pose significant therapeutic challenges. The emergence of programmed death-1 (PD-1) and programmed death-ligand 1 (PD-L1) checkpoint inhibitors has offered new clinical options; however, single-agent response rates remain modest—approximately 12–26% in selected populations 1112. Understanding why many patients fail to respond requires a detailed appreciation of how HPV-associated oncogenic signaling may contribute to an immunosuppressive tumor microenvironment (TME), including PD-L1 upregulation, regulatory T cell (Treg) infiltration, and myeloid-derived suppressor cell (MDSC) accumulation.
HPV Oncoprotein-Driven Immune Evasion
High-risk HPV encodes two principal oncoproteins, E6 and E7, which are constitutively expressed in cervical cancer cells at substantially higher levels than in other HPV-associated cancers. Cervical tumors can harbor viral loads ranging from fewer than 1 to 694 copies per cell, providing a sustained source of oncogenic and immune-suppressive signaling 5. The E6I splice isoform predominates among E6 splice variants in some HPV16-positive cervical tumors, but E6I should not be interpreted as preferential expression of full-length E6 protein 5.
E6 and E7 exert immune suppression through multiple convergent mechanisms. E6 directly inactivates p53, while E7 targets the retinoblastoma (Rb) protein, dismantling two pillars of innate immune signaling and type I interferon responses 4. Restoration of p53 and Rb function through siRNA-mediated E6/E7 silencing not only sensitizes cells to radiotherapy but also re-engages innate immune pathways, underscoring the centrality of these oncoproteins to immune evasion 4. Beyond tumor suppressor inactivation, E6, E7, and the HPV E2 protein directly enhance interleukin-10 (IL-10) gene transcription in infected keratinocytes, dendritic cells, tumor-associated macrophages, and Treg cells 10. IL-10, in turn, stimulates further E6 and E7 expression, establishing a self-amplifying positive feedback loop that persistently suppresses antigen-presenting cell function and Th1 polarization 10. Cervical IL-10 levels increase progressively during squamous intraepithelial lesion (SIL) development and peak within frank carcinoma 10, making this cytokine both a marker and mediator of progressive immune tolerance.
PD-L1 Upregulation: Transcriptional and Post-Transcriptional Regulation
A mechanistic study reported that HPV16 E6, rather than E7, promotes PD-L1 expression in HPV-positive cervical cancer models and in some lung cancer cell models 1. Knockdown of E6 in E6-positive cell lines (TL-1, SiHa) markedly decreased PD-L1 protein levels, whereas E6 overexpression in E6-negative lines increased PD-L1; E7 manipulation had no effect 1. Immunohistochemical analysis of 122 lung cancer specimens confirmed that combined PD-L1-positive/E6-positive tumors carried the worst prognosis (5-year OS 6.4%), illustrating clinical relevance beyond in vitro models 1. Functionally, experimental data suggest that E6-associated PD-L1 expression may contribute to colony formation and anchorage-independent growth, supporting a potential tumor-intrinsic role for PD-L1 in E6-associated oncogenic signaling 1.
Upstream signaling analysis indicates that PD-L1 induction is primarily mediated through MEK/ERK and NF-κB pathways, with selective kinase inhibitors (AZD6244 and BAY11-7082, respectively) suppressing PD-L1 expression, whereas PI3K/AKT inhibition had comparatively little effect 322. Additional transcriptional regulators include hypoxia-inducible factor-1 alpha (HIF-1α) and signal transducer and activator of transcription-3 (STAT3), both of which bind the PD-L1 promoter to amplify expression under inflammatory or hypoxic conditions 22. The interferon-inducible protein IFI16 further activates PD-L1 transcription through the STING-TBK1-NF-κB axis, which is particularly active in HPV-positive cells 31.
Post-transcriptional control adds important complexity. Loss of microRNAs (miRNAs) miR-140, miR-142, miR-340, and miR-383, which directly target PD-L1 mRNA, contributes to PD-L1 upregulation in cervical cancer 26. Conversely, miR-18a enhances PD-L1 indirectly by targeting the phosphatase PTEN, the ERK inhibitor WNK2, and the Wnt/β-catenin inhibitor SOX6, thereby activating PI3K/AKT, MEK/ERK, and Wnt/β-catenin pathways simultaneously 26. The OCT4/miR-18a axis links pluripotency signaling to immune checkpoint regulation. The miR-155 axis presents context-dependent complexity: while miR-155 is overexpressed in cervical cancer tissues and independently predicts poor prognosis (HR 2.320; 95% CI 1.259–4.276) 23, it also suppresses PD-L1 post-transcriptionally under TNF-α/IFN-γ stimulation via direct binding to PD-L1 3'-UTR sites 19, suggesting a temporal regulatory switch during tumor evolution. The STAT3-targeting miR-125a, suppressed through HPV-mediated p53 degradation, represents an additional layer of lost regulatory control 21.
| Regulatory Mechanism | Key Factors | Effect on PD-L1 | HPV Link |
|---|---|---|---|
| Transcriptional (signaling) | MEK/ERK, NF-κB, HIF-1α, STAT3 | Upregulation | E6 activates MEK/ERK and NF-κB 122 |
| Transcriptional (IFI16) | STING-TBK1-NF-κB | Upregulation | Active in HPV+ cells 31 |
| Post-transcriptional (miRNA loss) | miR-140, -142, -340, -383 | Upregulation (loss of suppression) | HPV disrupts miRNA networks 26 |
| Post-transcriptional (miR-18a/OCT4) | PTEN, WNK2, SOX6 targets | Upregulation (indirect) | OCT4/miR-18a-associated signaling in cervical cancer 26 |
| Post-transcriptional (miR-155) | PD-L1 3'-UTR direct targeting | Context-dependent suppression | Overexpressed but also suppressed by E6 1923 |
Regulatory T Cell Infiltration
Treg accumulation in the cervical cancer TME is driven by a coordinated chemokine and cytokine axis centered on CCL22 and its receptor CCR4. Quantitative mRNA analysis of cervical squamous cell carcinoma specimens versus normal cervix revealed markedly elevated FoxP3, CCL22, and CCR4 transcript levels, with all three factors showing strong positive pairwise correlations and further elevation in high-grade compared to low-grade tumors 17. Critically, in HPV-positive tissues, FoxP3 and CCL22 were upregulated while Smad3 and the OX40L/OX40 co-stimulatory axis were significantly downregulated, suggesting that HPV infection may be associated with a local immune milieu that favors Treg infiltration 17.
Once recruited, Tregs are sustained by IL-10 and transforming growth factor-beta (TGF-β). Flow cytometric analysis of peripheral blood in cervical cancer patients confirmed elevated PD-1 expression on CD4+CD25+FoxP3+ Tregs, with PD-1-high Tregs correlating positively with serum TGF-β and IL-10 levels and negatively with IFN-γ production 2. This PD-1/PD-L1 co-engagement on Tregs creates a self-reinforcing circuit: Treg-derived IL-10 and TGF-β further suppress effector CD8+ T cell function and simultaneously promote additional CCL22 production by macrophages and dendritic cells, recruiting more Tregs. PD-1 expression on CD8+ T cells additionally correlates with tumor differentiation status, lymph node metastasis, and distant metastasis, establishing PD-1+ CD8+ T cell frequency as a clinically relevant biomarker for advanced disease 2. In vitro PD-L1 blockade reduced IL-10 and TGF-β secretion and enhanced lymphocyte proliferation, confirming that checkpoint inhibition can partially reverse Treg-mediated suppression 2.
MDSC Accumulation and Suppressive Function
Tumor-derived granulocyte-colony stimulating factor (G-CSF) is a primary driver of MDSC (myeloid-derived suppressor cell) expansion in the cervical cancer TME. Cervical cancer cells engineered to express G-CSF accumulated significantly higher numbers of tumor-infiltrating CD11b+Gr-1+ MDSCs, predominantly of the granulocytic subtype (CD11b+Ly-6G+Ly-6Clow, >80% of CD11b+ infiltrate), compared to controls 18. Downstream, STAT3 and CCAAT/enhancer-binding protein beta (C/EBPβ) drive transcriptional programs promoting immature myeloid cell survival and impeding terminal differentiation 18.
A particularly consequential MDSC function involves prostaglandin E2 (PGE2) production. G-CSF-expanded MDSCs produce elevated PGE2, which signals through EP4 receptors on cervical cancer cells to promote aldehyde dehydrogenase (ALDH)-high cancer stem-like cell (CSC) properties—a phenotype associated with chemotherapy and radiotherapy resistance 18. Clinically, cervical cancer patients with elevated tumor-infiltrating MDSCs (CD11b+CD33+HLA-DR-) display significantly higher serum PGE2 concentrations and stronger ALDH1 immunoreactivity 18. Selective COX-2 inhibition with celecoxib suppressed PGE2-mediated CSC induction and enhanced cisplatin efficacy in preclinical models, without reducing MDSC frequency per se, identifying the MDSC-PGE2-CSC axis as a druggable resistance pathway 18. MDSCs also suppress T cell function through classical arginase-1, inducible nitric oxide synthase (iNOS), and reactive oxygen species (ROS) mechanisms, while IL-6 and IL-8 production further promotes angiogenesis and immunosuppression.
Implications for Checkpoint Inhibitor Therapy
The convergence of E6-driven PD-L1 expression, Treg accumulation, and MDSC expansion explains the modest efficacy of anti-PD-1/PD-L1 monotherapy in cervical cancer. KEYNOTE-158 established pembrolizumab's activity in PD-L1 CPS ≥1 disease (ORR 14.6%), with all responses occurring in this subgroup, leading to FDA accelerated approval 11. CheckMate 358 reported a higher ORR of 26.3% for nivolumab in a predominantly HPV-positive cohort, with durable responses (median not reached) and a 24-month OS of 49.8% 12. Strikingly, responses occurred in both PD-L1-positive and PD-L1-negative tumors in the nivolumab cohort, underscoring that PD-L1 expression alone is an imperfect predictive biomarker 12.
The most important mechanistic insight for combination therapy is that HPV-specific vaccine immunization induces adaptive PD-L1 upregulation as a resistance mechanism: Ad5-E6/E7 vaccination increased HPV-specific CD8+ tumor-infiltrating lymphocytes (TILs) but simultaneously induced compensatory PD-L1 on tumor cells and elevated PD-1+ TIL frequency. Addition of anti-PD-1 antibody reversed this adaptive resistance, maintaining CD8+ TIL expansion while reducing PD-L1 and PD-1+ TIL frequency, achieving superior tumor regression compared to either agent alone 9. Similarly, combining anti-PD-L1 monoclonal antibody with an Lm-LLO-E6 (Listeria-based) therapeutic vaccine produced superior tumor suppression and survival prolongation compared to monotherapy in both subcutaneous and metastatic HPV-positive tumor models 1.
The largest clinical advance has come from combination therapy. KEYNOTE-826 demonstrated that pembrolizumab plus chemotherapy with or without bevacizumab in first-line PD-L1 CPS ≥1 disease produced a median OS not reached versus 16.3 months (HR 0.64), median PFS of 10.4 versus 8.2 months (HR 0.62), and ORR of 68% versus 50% 32. EMPOWER-Cervical 1 established cemiplimab monotherapy versus single-agent chemotherapy in the second-line setting, demonstrating a median OS benefit of 12.0 versus 8.5 months (HR 0.69) across all histologies, with consistent benefit in both squamous cell carcinoma (HR 0.73) and adenocarcinoma/adenosquamous carcinoma (HR 0.56) 33. Notably, cemiplimab responses occurred in both PD-L1-positive and PD-L1-negative patients 33, further challenging PD-L1 CPS as the sole determinant of therapy selection. Pembrolizumab has an FDA tissue-agnostic indication for unresectable or metastatic TMB-high solid tumors (TMB ≥10 mutations/Mb) that have progressed after prior treatment and have no satisfactory alternative treatment options; the specific predictive value of TMB in HPV-driven cervical cancer remains uncertain and warrants further study 34.
Future Directions and Conclusions
A rational framework for improved immunotherapy in cervical cancer emerges from the mechanistic evidence reviewed here. Anti-PD-L1 combined with CCL22/CCR4 axis disruption or TGF-β blockade could simultaneously relieve checkpoint inhibition and reduce Treg suppression. MDSC-targeting via COX-2 inhibition (celecoxib) combined with checkpoint blockade addresses PGE2-driven stemness and treatment resistance. Dual checkpoint blockade (PD-1 + CTLA-4) and HPV antigen-directed therapeutic vaccines combined with checkpoint inhibitors address the adaptive immune resistance induced by vaccine-driven immune activation. Future biomarker strategies should integrate PD-L1 CPS, TMB, FoxP3/CCL22 expression, circulating MDSC burden, serum PGE2, and HPV E6/E7 mRNA quantification—the latter demonstrating an odds ratio of 8.97 for high-grade disease 6—to enable precision patient stratification in prospective immunotherapy trials.
In summary, cervical cancer's resistance to single-agent checkpoint inhibition reflects HPV E6-driven PD-L1 upregulation through MEK/ERK and NF-κB signaling reinforced by miRNA dysregulation, Treg accumulation orchestrated by the CCL22/CCR4 axis and sustained by IL-10/TGF-β, and MDSC-derived PGE2 that promotes cancer stemness and therapeutic resistance. Targeting these interconnected immunosuppressive pathways through rational combination strategies represents a scientifically plausible approach to improving outcomes in this disease.