Abstract
Background:
Preclinical data suggests a potential role for OX40 agonists in cancer treatment, though clinical trials have yielded modest results.
Objectives:
We assessed OX40 and other immune checkpoint RNA expression in advanced solid malignancies and their clinical outcomes.
Design:
We conducted a retrospective cohort study analyzing tissue samples from 514 cancer patients.
Methods:
Immune marker transcripts were evaluated using a clinical-grade laboratory test. OX40 expression was normalized, and percentiles ranked as “moderate/low” (0–74) or “high” (75–100).
Results:
High OX40 expression was found in 24% of tumors, primarily in esophageal, liver/bile duct, stomach, small intestine, and lung cancers. High OX40 was associated with increased expression of programmed death-1, cytotoxic T-lymphocyte-associated antigen-4, and forkhead box P3 but did not predict overall survival in immunotherapy-naïve patients or outcomes post-immunotherapy.
Conclusion:
OX40 expression is heterogeneous and strongly associated with immunosuppressive checkpoints, potentially suggesting that effective use of OX40 agonists may depend on patient selection based on tumor immunomics.
Plain language summary
OX40 is a small protein on activated T cells that acts as an “on switch booster,” helping these immune cells fight infections and cancers. This study of 514 patients with advanced cancers revealed that about one-quarter had high OX40 expression. Interestingly, tumors with high levels of OX40 also exhibited elevated levels of other signals (such as PD-1, CTLA-4, and FOXP3) that function as “brakes” on the immune system. This likely suggests that these tumors are in a mixed state, possessing both “go” and “stop” signals. Importantly, increased OX40 did not correlate with longer patient survival, whether or not they received immunotherapy. This suggests that simply activating OX40 may not be enough, and future treatments should focus on carefully selecting patients based on their tumor biology. Notably, recent data shows that the OX40 inhibitor was withdrawn from the market because it is linked to Kaposi’s sarcoma, a cancer that appears as purple, red, or brown patches on the skin and starts in the cells lining small blood and lymph vessels.
Introduction
OX40 co-stimulatory receptor (CD134) and ligand (CD252) belong to the tumor necrosis factor receptor superfamily (TNFRSF) and are also known as TNFRSF4 and TNFSF4, respectively. OX40 is expressed by activated T cells, which results in T-cell survival, proliferation, and regulatory T-cell inhibition via decreased transcription of genes regulating cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4) and forkhead box P3 (FOXP3).1–3 The tumor microenvironment (TME) is complex, and immune evasion often involves multiple immune markers, such as lymphocyte-activation gene 3 (LAG-3), which drives T-cell exhaustion and is associated with the expression of other immune checkpoints, including program death-1 (PD-1), program death ligand-1 (PD-L1), and CTLA-4.4,5 A phase II–III clinical trial of relatlimab targeting LAG-3 plus nivolumab demonstrated improved outcomes compared to nivolumab monotherapy in patients with metastatic melanoma resulting in Food and Drug Administration (FDA) approval. 6 However, despite compelling preclinical data, early-phase clinical studies of OX40 agonists have shown limited activity, and our prior work suggests that only a subset of patients with advanced solid tumors (~17%) exhibit high OX40 but low OX40L expression, a pattern that may be biologically amenable to OX40 agonist therapy. 3 In parallel immune checkpoint inhibitors (ICIs) targeting PD-1, PD-L1, CTLA-4, and LAG-3 have transformed the treatment landscape for a diverse solid malignancy. Microsatellite instability-high (MSI-H), high tumor mutational burden (TMB), and high PD-L1 expression are well-recognized immune biomarkers for response to ICIs.7–9 For example, pembrolizumab is FDA-approved for the treatment of advanced solid tumor malignancies based on biomarkers, MSI-H, and/or high TMB (⩾10 mutations/megabase).8,10 Other biomarkers for response may include PD-1 expression on tumor-infiltrating lymphocytes. 11 Interestingly, very high TMB (⩾50 mutations/megabase) may also be a prognostic marker associated with improved survival in immunotherapy-naïve cancer patients, likely due to endogenous immune mechanisms. 12 A recent meta-analysis of over 19,000 patients found that most immune-oncology trials did not use biomarkers for patient selection, despite evidence that these biomarkers are linked to better immunotherapy outcomes. 13
Given the disconnect between the promising preclinical hypotheses regarding OX40 and the limited efficacy observed in clinical trial settings, it is essential to gain a better understanding of how T cells interact with the TME through co-stimulatory receptors and immune checkpoints. Therefore, this study aims to evaluate the RNA transcriptomic expression of OX40 and OX40L across various advanced solid tumors, analyze their associations with other immune markers, and explore their impact on clinical outcomes.
Materials and methods
This study is a retrospective cohort analysis of patients with advanced solid tumors, adhering to the STROBE guidelines for observational studies (https://www.equator-network.org/reporting-guidelines/strobe/, Supplemental File-STROBE checklist).
Patient selection
We analyzed 514 advanced solid tumor patients (UCSD Moores Cancer Center) who underwent tissue RNA sequencing (CLIA-certified, CAP-accredited laboratory; Labcorp, Buffalo, NY, USA; formerly OmniSeq). The study followed the Institutional Review Board-approved UCSD PREDICT protocol (NCT02478931) with patient consent for investigational procedures.
Tissue sampling and immune marker analysis
RNA was extracted from formalin-fixed, paraffin-embedded tumors (truXTRAC® FFPE kit (Covaris, Woburn, MA, USA)) and quantified with Quant-iT RNA HS assays (Thermo Fisher Scientific, Waltham, MA, USA). Ex Torrent Suite’s immuneResponseRNA plugin (v5.2.0.0), generated expression counts followed by normalization and percentile ranking. An internal housekeeping profile data set was used to assess transcript abundance, normalized and ranked on a 0–100 scale based on a reference of 735 tumors from 35 malignancies. OX40 expression was classified as Moderate/Low (0–74) or High (75–100). This reference cohort of pretreated solid tumors shows diverse immune expression. Using the ⩾75th percentile as a cutoff, we defined high expression as the top quartile, providing a practical though somewhat arbitrary threshold unless distributions are markedly skewed (Supplemental References). PD-L1 was assessed by immunohistochemistry.
Outcomes and statistical analyses
Group comparisons used t-tests for continuous variables; Fisher’s exact tests for categorical variables. Survival outcomes were analyzed using the Kaplan–Meier method. Overall survival (OS) was measured from advanced disease diagnosis to death or last follow-up (June 24, 2022). For immunotherapy-treated patients, OS was calculated from treatment initiation; progression-free survival (PFS) was calculated from treatment start to progression or censoring. Significant variables from univariate analyses were included in multivariate logistic regression. Analyses were performed using R 4.2.1, with significance defined as two-sided p ⩽ 0.05. The database for the materials used in this article can be made available to investigators upon request.
Results
Patient characteristics
Altogether, 514 patients (489 with advanced/metastatic disease and clinical outcome data) with different histological diagnoses were included for analysis in this study. The median age of diagnosis of metastatic/locally advanced disease was 61 years; 40% (240) were men. The most common tumor types were colorectal (N = 140), pancreatic (N = 55), breast (N = 49), and ovarian cancer (N = 43; Table 1). Of 217 patients (42%) who received ICIs, all but two patients received an anti-PD-1 or anti-PD-L1 agent; two patients received ipilimumab monotherapy.
Patient characteristics and other immune markers.
CTLA-4, cytotoxic T-lymphocyte associated protein 4; FOXP3, forkhead box P3; ICOS, inducible T-cell costimulatory receptor; IHC, immunohistochemistry; LAG-3, lymphocyte-activation gene 3; Mb, megabase; MSI, microsatellite instability; muts, mutations; PD-1, program death-1; PD-L1, program death-ligand 1; TIGIT, T cell immunoreceptor with immunoglobulin and ITIM domain; TIM3, T-cell immunoglobulin and mucin domain 3; TMB, tumor mutational burden; TNFSF18, tumor necrosis factor ligand superfamily member 18.
OX40 transcript expression varied within and between malignancies
We interrogated OX40 RNA expression across diverse malignancies (N = 514). Overall, approximately 24% of tumor samples (122 of 514) had high OX40 RNA expression (⩾75 rank; Figure 1(a)). Table 1 outlines the characteristics of patients with high OX40 expression or moderate/low OX40 expression. The median age of diagnosis, sex distribution, PD-L1 expression by IHC, and high TMB were similar in both groups.
Distribution of OX40 RNA expression, tumor-type variability, and immune correlates. (a) Histogram of OX40 RNA expression levels. Each number above a bar indicates the number of patients who had an RNA rank score within the range of that bar. Overall, 122 patients (24% (122 of 514)) had an RNA rank score of 75 or more. (b) Expression patterns of OX40 across cancer types. In each row, red indicates high expression (⩾75), gray indicates moderate expression (25–74), and blue indicates low expression (<24). Cancer types are sorted according to the proportion of patients who had high OX40 expression. The highest proportion of high OX40 expression was observed in esophageal cancer (41%), followed by liver and bile duct cancer (37%), stomach cancer (36%), small intestine cancer (33%), lung cancer (30%), and pancreatic cancer (27%). (c) Multivariable logistic regression analysis for the association of high OX40 RNA expression and other immune markers. High expression refers to ⩾75th RNA percentile rank. The figure shows that high OX40 expression is independently associated with high PD-1 expression, high CTLA-4 expression, and high FOXP3 expression.
The tumor types with the highest OX40 RNA expression were esophageal (41%), liver and bile duct (37%), stomach (36%), small intestine (33%), and lung (30%), respectively. Uterine (8%), head and neck (8%), neuroendocrine (13%), and ovarian cancers (16%) were observed to have a low frequency of high OX40 transcript expression. Figure 1(b) shows variable OX40 expression across different malignancies.
High transcriptomic levels of OX40 correlated with other checkpoints (PD-1, CTLA-4, FOXP3)
On univariable analysis, high OX40 expression was associated with high RNA transcriptomic expression of inhibitory immune checkpoints for T-cell recognition (PD-1, PD-L1, CTLA-4, LAG-3, TIM-3, TIGIT), co-stimulatory receptors (CD137/4-1BB, ICOS, CD226), and T-regulatory cell marker (FOXP3; Table 1). No significant correlation was found between high OX40 expression and OX40L RNA expression.
On multivariable logistic regression analysis, high OX40 RNA expression was independently associated with high PD-1 RNA expression (odds ratio (OR) 2.92, 95% confidence interval (CI) 1.43–5.93, p = 0.003), high RNA expression of CTLA-4 (OR 3.87, 95% CI 1.91–7.96, p < 0.001), and high RNA FOXP3 expression (OR 3.04, 95% CI 1.67–5.49, p < 0.001; Figure 1(c)).
OX40 RNA expression is not a prognostic factor for survival in patients with advanced/metastatic cancer
Survival was explored in patients whose tumors had high OX40 RNA expression (75–100th percentile RNA rank) versus moderate/low OX40 RNA expression (0–74th percentile RNA rank). In the entire cohort of 489 patients with advanced/metastatic cancer and clinically available data, as well as in the 272 patients who were immunotherapy naïve, high versus not-high OX40 RNA levels were not predictive of OS from the time of diagnosis of advanced/metastatic disease (Figure 2(a) and (b); p = 0.45 and p = 0.98, respectively).
Survival analysis stratified by OX40 RNA expression levels. The percentile of the OX40 expression based on RNA level in each tumor was ranked on a scale of 1–100 and ranked into moderate/low (0–74) and high (75–100). No difference in OS from the time of diagnosis of advanced/metastatic disease was observed in the entire cohort (a) and immunotherapy-naïve patients (b) when examining OX40 high expression versus moderate/low OX40 expression. In patients with received immune checkpoint blockade, progression-free survival (c) and OS (d) from the time of start of the first immunotherapy) did not segregate by OX40 levels. The log-rank test was performed to analyze survival data.
OX40 RNA expression is not a predictive factor for progression-free or OS after immunotherapy in patients with advanced/metastatic cancer
Outcomes were evaluated in 217 patients who received an ICI. Patients whose tumors had high (>75th percentile RNA rank) versus moderate/low (<75th percentile RNA rank) OX40 expression did not differ in their PFS or OS from the date of initiation with the first ICI (p = 0.82 and p = 0.16, respectively; Figure 2(c) and (d)).
OX40 together with OX40L expression is not a predictive indicator for outcome
The survival outcome was further explored based on the combination of OX40 and OX40L expression, which was divided into four categories: high OX40 (75–100 percentile) and moderate/low OX40L (0–74 percentile), high OX40 and high OX40L, moderate/low OX40, and moderate/low OX40L, and moderate/low OX40 and high OX40L.
OX40 and OX40L expression in any combination examined was not associated with OS in the entire cohort of patients or in the immunotherapy-naïve subset of patients. Nor was any combination of OX40 and OX40L expression correlated with PFS or OS after ICI treatment (Supplemental Figure 1).
Discussion
Our study presents important findings, particularly the absence of a relationship between OX40 expression and clinical outcomes (OS/PFS), regardless of the ICI treatment received. Particularly, high OX40 expression was associated with high RNA expression levels of immunosuppressive checkpoints (PD-1, CTLA-4, FOXP3). We found high OX40 RNA expression in about 24% of diverse cancers, notably in esophageal, liver and bile duct, stomach, small intestine, and lung cancer, though expression levels varied. CTLA-4 and PD-1 maintain immune tolerance, while FOXP3, expressed by T-regulatory cells, contributes to their immunosuppressive potential.14,15 Thus, high OX40 expression may indicate an immunosuppressive environment. Previous studies showed that T-regulatory cells in hepatocellular carcinoma (HCC) often have high OX40 levels alongside markers like LAG3, PD-1, TIM-3, CD8, and CD68 expression. 16 Taking these observations together, it may be inferred that interrogating each tumor’s immunomic status could be important when crafting therapy with OX40 agonists, especially when combined with ICIs. Failure to do so may underlie the relative lack of clinical salutary effects (so far) of OX40 agonists. 3
In our analysis, OX40 and OX40L expression did not correlate with cancer patient survival or PFS/OS after ICI treatment. This lack of prognostic value may arise from our evaluation across diverse cancers. Additionally, we lack data on OX40 protein expression, which could provide further insights. To better comprehend OX40 as a potential biomarker, more extensive translational research integrating multigene immune signatures with spatial immune phenotype mapping in the TME is needed. Previous studies have yielded inconsistent results. For instance, Xie et al. 16 noted that high OX40 IHC expression correlated with poor survival for patients with HCC. On the other hand, in patients with colorectal cancer, high infiltration of tumor tissue by CD8+ T cells with high OX40 gene expression correlated with better outcomes. 17 It is important to note that in Kaposi sarcoma (KS), OX40 is mainly an inducible receptor on activated T cells, while its ligand OX40L is strongly expressed on HHV-8-infected endothelial cells in lesions. OX40 deficiency, an autosomal-recessive genetic condition, causes severe CD4 memory dysfunction and is a known cause of childhood-onset classic KS, underscoring its critical role in controlling HHV-8 in endothelial tissues. 18 Interestingly, recent data indicate that the OX40 inhibitor (Rocatinlimab), currently in late-stage development for atopic dermatitis, 19 has been withdrawn from the market due to its association with causing KS (https://www.dermatologytimes.com/view/kyowa-kirin-ends-late-stage-ox40-development).
Our study has several limitations, including the real-world nature of the data as well as the lack of information on OX40 expression in specific cell types. We also have limited data on individual histologies, an issue that merits future investigation. Finally, we focused on transcriptomic data, without data on protein levels.
Conclusion
We found that transcriptomic expression of OX40 was heterogeneous within the same tumor type and across different malignancies. Moreover, high OX40 expression significantly and independently correlated with CTLA-4, PD-1, and FOXP3 transcriptomic expression. Therefore, identifying tumor-specific expression patterns and co-targeting other expressed immune checkpoints warrants investigation as a precision immunotherapy strategy.
Supplemental Material
sj-docx-1-tam-10.1177_17588359261456325 – Supplemental material for OX40 transcriptomic expression and its association with immune checkpoints and clinical outcomes
Supplemental material, sj-docx-1-tam-10.1177_17588359261456325 for OX40 transcriptomic expression and its association with immune checkpoints and clinical outcomes by Bicky Thapa, Daisuke Nishizaki, Hirotaka Miyashita, Suzanna Lee, Mary K. Nesline, Rebecca A. Previs, Jeffrey M. Conroy, Taylor J. Jensen, Paul DePietro, Sarabjot Pabla, Shumei Kato and Razelle Kurzrock in Therapeutic Advances in Medical Oncology
Footnotes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
