Research Highlights: Immunotherapy to Overcome Lung Tumor Cell-Induced Escape from Immunosurveillance

The activation of immune responses to tumors in cancer patients has become well established; hence, tumor eradiation through antigen-based immunobiotherapy is theoretically possible. However, vaccine trials and other forms of immune augmentation in solid tumors have been disappointing, owing in large part to the ability of tumor cells to circumvent and suppress immune activation, expansion and effector functions. A number of different immune mechanisms that regulate antigenic responses have been usurped by tumor cells to ambush both antigen-presenting dendritic cells and immune lymphoid cells that migrate into tumor microenvironments. The three articles chosen for critical analyses present one mechanism through which tumor cells escape immune surveillance and data from a clinical trail attempting to overcome that mechanism of immune suppression. Specifically, the drug used in the clinical trial targets the interaction between a cell surface protein on T cells whose function normally serves as a self-regulator of T-cell expansion, programmed cell death receptor-1 (PD-1; CD279), and its ligands. In normal tissue this receptor–ligand interaction prevents the immune system from being overwhelmed with antigenically stimulated T cells that have served their functions. Negative regulatory signals have thus evolved to control T-cell expansion after an antigenic response clears infectious organisms from the body by stimulating the maturation of a subset of responsive cells into memory T cells and discarding the remainder. PD-1 interacts with its ligand PD-L1 (B7-H1, CD274) to negatively regulate activated T cells by limiting their expansion and cytokine production, and then by inducing programmed cell death (reviewed in [1–3]). However, some tumors, particularly of the epithelial type, have upregulated their cellular expression of ligands for PD-1 in order to survive and, as such, are able to turn off effector T cells that might otherwise invade the tumor microenvironment and kill tumor cells. Non-small-cell lung cancer (NSCLC) is one of several tumors where investigators have demonstrated significant aberrantly upregulated expression levels of two PD-1 ligands – PD-L1 and PD-L2 (B7-DC, CD273) – on tumor cell membranes [4].

Evaluation of: Mu CY, Huang JA, Chen Y, Chen C, Zhang XG: High Expression of PD-L1 in Lung Cancer May Contribute to Poor Prognosis and Tumor Cells Immune Escape Through Suppressing Tumor Infiltrating Dendritic Cells Maturation. Med. Oncol. DOI: 10.1007/s12032-010-9515-2 (2010) (Epub Ahead of Print).

In a recent publication by Mu et al., the expression of the ligand PD-L1 and its relationship to tumor-infiltrating dendritic cells was investigated in 129 NSCLC surgically resected tissues [5]. Patients were equally distributed among stages I–III, and 42% were adenocarcinoma subtype. First, the article showed by immunohistochemistry that PD-L1 expression was present in 53% of cancer cells and only in approximately 5% of adjacent lung parenchyma. More PD-L1-positive cells were seen in adenocarcinoma compared with squamous tumors (65.2 vs 44.4%; p = 0.032). Second, a significant correlation was revealed between the prognosis and the level of expression of PD-L1 on tumor cells. Higher levels of PD-L1 expression in these lung cancer patients was associated with shorter survival periods, with a p-value of 0.034. This relationship differs from that presented in a previous paper [4], perhaps because Mu et al. examined a larger group of patients. Mu et al. used two markers and morphologic appearances to characterize tumor-infiltrating dendritic cells as immature versus mature based on CD1α and CD83 expression, respectively [5]. The presence of immature CD1α⁺ dendritic cells within tumors has been associated with better prognosis [6]. Combining marker detection with morphology allowed Mu et al. to report a more accurate assessment of quantities of dendritic cells, since CD1α and CD83 markers are not exclusively expressed on dendritic cells. A correlation of immaturity (CD1α expression) of tumor-infiltrating dendritic cells and increased PD-L1 expression on tumors cells was reported. PD-L1 may be turning off the antigen-presenting abilities and maturation of dendritic cells, hence decreasing their chances to enhance tumor-induced bronchus-associated lymphoid tissue (Ti-BALT) formation in lung tumors and specific immune responses developed therein. The localization of mature dendritic cells, identified with the marker DC-Lamp, into Ti-BALT was highly associated with a favorable clinical outcome following surgical resection of lung tumors in another study of mature dendritic cells [7]. A finding of PD-L1 expression on tumor cells and increased dendritic cell immaturity (less CD45 expression) was also seen in previous work by Konishi et al., again suggesting a possible link between PD-L1 and dendritic cell maturation [4]. Migration of dendritic or lymphoid cells into lung tumors was characterized in each paper, and associations with PD-L1 expression lead to the conclusion that PD-L1 plays an active role in undermining patients‘ immune responses to tumor cells. Mu et al. also presented confocal micrograms which demonstrated an overlap of CD1α and PD-L1 expression on the same cells [5], although these double-positive dendritic cells may in fact be epithelial cells in the tumor microenvironment [8]. Future research should clearly focus on determining the identities of double-positive CD1α⁺, PD-L1⁺ cells. Furthermore, the localization of mature and/or immature dendritic cells to Ti-BALTs in PD-L1⁺ tumor microenvironments needs to be characterized. The phenotypic characterization of immune cells in tumors from surgically resected lung cancer patients had previously demonstrated the presence of mature dendritic cells and mature, granzyme⁺, presumably CD8⁺ effector cells in Ti-BALTs [7].

Evaluation of: Zhang Y, Huang S, Gong D, Qin Y, Shen Q: Programmed Death-1 up Regulation is Correlated with Dysfunction of Tumor-Infiltrating CD8+ T Lymphocytes in Human Non-Small Cell Lung Cancer. Cell. Mol. Immunol. DOI: 10.1038/cmi.2010.28 (2010) (Epub Ahead of Print).

In a second recent publication on the PD-1–PD-L1 pathway, Zhang et al. phenotypically characterized immune cells in NSCLC tumor microenvironments and compared them with circulating CD8⁺ cells in lung cancer patients and in normal donors [9]. The study demonstrated that approximately 50% of CD8⁺ T cells isolated from tumors express high levels of PD-1, the receptor for PD-L1. In examining peripheral blood lymphocytes, CD8⁺ cells in NSCLC patients had a significantly higher percentage of PD-1 expression compared with CD8⁺ T cells from normal donors. The intensity of staining for PD-1 – the quantity of surface expression – was significantly greater on CD8⁺ T cells within tumors as opposed to the level expressed on cancer patients‘ own peripheral blood CD8⁺ cells. Functional analyses revealed that CD8⁺ cells isolated from tumors had decreased proliferative responses and decreased cytokine production compared with CD8⁺ cells from peripheral blood. The investigators attributed this functional exhaustion of CD8⁺ T cells within the tumors to PD-1 expression on the tumor-infiltrating cells. Unfortunately, the investigators did not stain the CD8⁺ cells for all three markers – with anti-PD-1 as well as anti-CD8 and anti-IFN-γ – such that one could definitely verify that it was the PD-1⁺, CD8⁺ population that displayed decreased cytokine synthesis. Triple staining of the activated cells would have allowed for a comparison of cytokine expression in CD8⁺ cells that lacked PD-1 expression with those that were PD-1⁺. Experiments where PD-1 receptor–ligand interactions were blocked with antibodies to PD-1 ligands revealed that disrupting PD-1–PD-L1 interactions led to increased immune functions, including IFN-γ production and proliferation of CD8⁺ T cells, although these responses were still far below responses of CD8⁺ T cells isolated from patients‘ peripheral blood. The failure to reach normal levels of proliferation or cytokine production may be a result of PD-1 pathway induction of apoptosis, which was not overcome in this study by blocking PD-1–PD-L1 interactions [9]. These studies suggest that anti-PD-1 immunotherapy in NSCLC cancer patients may be of benefit to prevent immune exhaustion of circulating CD8⁺ T cells as they migrate into tumors. Furthermore, they suggest that anti-PD-1 therapy would not significantly augment immune functions of immune cells present within the tumor microenvironment on day 1 of therapy, as they may be destined to die by apoptosis owing to their prior interaction with PD-L1 expressed on the tumor cells.

Evaluation of: Brahmer JR, Drake CG, Wollner I et al.: Phase I Study of Single-Agent Anti-Programmed Death-1 (MDX-1106) in Refractory Solid Tumors: Safety, Clinical Activity, Pharmacodynamics, and Immunologic Correlates. J. Clin. Oncol. 28, 3167–3175 (2010).

From the test tube demonstrating the potential of anti-PD-1 to inhibit PD-L1-triggered immunosuppression induced by tumor cells, we move to the bedside in the third paper selected for critical review. A Phase I clinical trial of antibody dosing with a fully human anti-PD-1 antibody (MDX-1106) was recently reported by Brahmer et al.[10]. Six NSCLC patients were included in this trial. Eligible patients had treatment-refractory disease. Patient cohorts received dose escalations of the anti-PD-1 antibody until a maximum tolerated dose was achieved. Although toxicity was the primary end point of this study, durable responses were observed in some patients, including one with NSCLC whose lesions regressed. Stable disease or lesional tumor regression was achieved in 12 out of 39 patients treated with anti-PD-1. Nine patients in the trial had tumor tissues stained for expression of the PD-1 ligand, B7-H1 (also known as PD-L1). The purpose of staining for B7-H1 was to distinguish patients whose tumor microenvironment revealed upregulated expression and to determine whether outcomes differed from patients with little or no PD-L1 expression. Four out of nine patients demonstrated upregulated expression of PD-L1 (B7-H1). Positive clinical responses were observed in three out of the four patients whose tumors demonstrated PD-L1 expression, while five out of five patients whose tumors did not express PD-L1 demonstrated no clinical responses [10]. Although these data represent a relatively small sample size, the correlation between PD-L1 expression of tumor cells and the likelihood of tumor regression following anti-PD-1 immunotherapy reached significance with a p-value of 0.0476. The information published in this article dramatically supports the need to select subsets of patients for immunotherapy. In future trials, lack of evidence for upregulated expression of PD-L1 (B7-H1) would be a rational exclusion criterion for a randomized clinical trial. Only patients with detectable PD-1 ligand levels, specifically PD-L1 (B7-H1), in their tumor microenvironments should be included in an immunotherapy Phase II or III trial with anti-PD-1.

Conclusion

Taken together, these three recent publications suggest that blocking negative immune-regulatory pathways expressed in T cells migrating to tumor microenvironments can enhance patients‘ immune responses to their tumors. These articles focused specifically on the PD-1–PD-L1 pathway, which downregulates immune functions, as evidenced by decreased immune cell proliferation and cytokine production, and increased immune cell apoptosis. Furthermore, patients whose tumor microenvironments demonstrate upregulated PD-1 ligand expression have a worse prognosis and may benefit from immunotherapy that blocks the PD-1 signaling pathways. The articles are in agreement that approximately 50% of NSCLC patients demonstrate upregulated expression of the ligand PD-L1 [4,5,10]. Characterizing the tumor microenvironment of patients is, therefore, of utmost importance in designing clinical trials for targeted immunotherapy. Treatment with anti-PD-1-targeted therapy in patients without PD-L1 expression in tumor microenvironments would probably not produce clinically significant benefits and be a waste of reagents, resources, time and financial investments, as well as increasing medical costs.

PD-1 signaling is complicated by the fact that tumor and stromal cells may contribute to the capacity to trigger this receptor as the ligand appears to be expressed on tumor cells, normal epithelial cells and immature dendritic cells, while the receptor PD-1 itself is upregulated on immune effector CD8⁺ cells within the tumor microenvironment. Immature dendritic cells are important in the initiation phase of immune responses to tumors, while the CD8⁺ T cells are important in the effector phase. The fact that antibodies that prevent PD-1–PD-L1 signaling do not allow immune cells found within the tumors to fully regain their functions suggests that the effectiveness of this immunotherapy will require the immigration of naive or activated circulating immune cells into the tumor environments. Secondary biopsies of primary or metastatic tumors should be performed on patients in a clinical trial, as it would allow investigators to determine whether anti-PD-1 therapy results in increased immune cell migration into the tumor. Combination therapies should also be carefully considered to optimize the effectiveness of both reagents. Some chemotherapeutic drugs have been shown to downregulate memory T cells and increase circulating naive cells, which might be considered as rational strategies to combine with anti-PD-1 and other targeted immunotherapies [11].

Financial & competing interests disclosure

Janet MD Plate is supported in part by the Wadsworth Foundation. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.

Additional information

Funding

Janet MD Plate is supported in part by the Wadsworth Foundation. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. No writing assistance was utilized in the production of this manuscript.