Research from a team at the University of Texas M.D. Anderson Cancer Center has revealed a number of different mechanisms that help explain why immunotherapy clinical trials targeting osteosarcoma have, thus far, failed to yield significant advances and suggested new therapeutic strategies based on these findings.
Osteosarcoma is the most common solid tumor arising from the bone. One of the hallmarks of this cancer is the presence of multiple genomic mutations, which would suggest therapies such as immune checkpoint inhibitors as an appealing therapeutic approach. Now, in research published in Nature Communications, the M.D. Anderson team revealed that poor infiltration of the tumor by immune cells, low activity from available T cells, a lack of immune-stimulating neoantigens, and multiple immune-suppressing pathways all combine to hinder response to immunotherapy.
When detected early in the course of disease, standard chemotherapy and surgical treatment for osteosarcoma is fairly effective with a 70% survival rate. But metastatic osteosarcoma is associated with survival rates between 20% and 30%. Clinical trial research with immunotherapy has been a focal point for these poor-risk patients, i.e. those with metastatic or relapsed osteosarcoma.
Immune Profiling
The team sought to understand the genetic/immunogenic profiles of a variety of osteosarcoma cases to learn more about the potential for immunotherapy. Their analysis included 48 poor-risk pediatric and adult patients with high-grade osteosarcoma (73 percent having a poor response to standard treatment). Over half (51 percent) of the samples in the study were from metastatic cancers. Another 23 percent were from relapsed patients. The remainder were primary osteosarcoma diagnoses.
The team performed whole-genome, RNA, and T-cell receptor sequencing, immunohistochemistry, and reverse phase protein array profiling to create an immune profile of each sample.
The whole-genome landscape confirmed earlier research by others, and there was little difference in genomic mutations between samples from primary, local recurrence, or metastatic lung specimens. The authors had suspected that the substantial burden of genomic mutations might yield an increase in mutated proteins – or neoantigens – which are considered a potential immune system trigger against tumor cells. This kind of association has been found in other cancer types, but was lacking in this analysis.
Poor T cell Infiltration
The team compared the immune infiltration by T cells in the osteosarcoma samples with those of other cancers profiled in The Cancer Genome Atlas (TCGA). They discovered that most samples contained medium or high levels of T cells (64 percent). However, the level of immune cell infiltration by T cells was lower than expected compared with other cancers, like lung cancer or melanoma, which respond more favorably to checkpoint inhibition. Further, the T cell subpopulations, or clones, lacked sufficient diversity to mount an effective immune response.
Their analysis suggested that only a minority of osteosarcoma samples from this study and other osteosarcoma tissue collections, including those from International Caner Genome Consortium (ICGC), and the Therapeutically Applicable Research to General Effective Treatments (TARGET) studies, may have sufficient T cell infiltration to elicit a response from checkpoint inhibition. In their sample, only 8 percent showed sufficient immune cell infiltrations compared with 10 percent in the ICGC, and 15 percent in the TARGET study.
Gene Expression Variation
Gene expression analysis revealed three distinct clusters of samples, corresponding with levels of immune infiltration. “Hot” tumors had the highest degree of immune infiltration, but also had high activity in a number of signaling pathways, including PD1, CTLA4, and IFNG, that inhibit T-cell activation.
Conversely, “cold” tumors had the lowest levels of immune infiltration, decreased expression of human leukocyte antigen (HLA) – an important molecule for communicating with immune cells – and a higher number of genes with copy number loss. They found that of the multiple genomic alterations in osteosarcoma samples, copy number loss was associated with the highest association with immune suppression.
Their analysis also revealed increased expression of the gene PARP2 in osteosarcoma samples.
Potential Therapeutic Strategies
Having defined three broad immune subsets of osteosarcoma samples corresponding with low, intermediate, and high levels of T cell immune infiltration, the authors stress the need to dive deeper into studying the multiple contributors to immune suppression in osteosarcoma.
“The immunogenomic landscape in osteosarcoma is characterized by genomic complexity and significant disease heterogeneity,” write the authors. “This complexity may contribute to an immunosuppressive phenotype through multiple mechanisms that may present themselves as opportunities for novel therapeutic exploitation.”
They suggest several approaches for future clinical and translational studies:
- Include osteosarcoma patients with high immune scores who may be more likely to benefit from checkpoint inhibition
- Explore new approaches to enhancing neoantigen expression via radiation, NMD inhibition, or intratumoral vaccines
- Develop combinations of targeted and immunotherapy. In patients intermediate to high immune infiltration, these might include checkpoint inhibition combined with targeted therapies to counterbalance the effects of immunosuppressive genes such as TGFb or PI3K. In low immune infiltration “cold” osteosarcomas, the authors suggest that increased PARP2 expression could be targeted by available PARP inhibitors, and combined with checkpoint blockade.