NEW YORK – A proof-of-principle study by the Leiden University Medical Center (LUMC) and the Biotech Support Group showed that a protein-based liquid biopsy can provide cancer prognostic information based on the tumor microenvironment.
The proteomic Stroma Liquid Biopsy (SLB), which the collaborators hope to eventually commercialize, estimates the risk of cancer progression by measuring 13 proteins largely related to innate immunity, as well as to inflammation and coagulation.
The collaborators recently published an explorative study that integrated single-cell and bulk transcriptomic data to characterize the SLB panel genes as part of an in silico proof of concept. In addition to providing a theoretical framework for using proteomic signatures to report on tumor-stroma content in a liquid biopsy, they identified a potential colon cancer survival biosignature.
Although the SLB is currently at an early developmental stage, the technique is novel for reversing the usual path to biomarker discovery.
Rather than building a gene expression signature from differentially expressed genes predictive of survival or other patient outcomes, the SLB panel arose from unbiased protein screens conducted by the Biotech Support Group in 2017, as part of an effort to evaluate the company's products.
BSG's sample prep methods include a less conventional "negative selection" process, in which the majority of the proteome is captured on beads, while allowing albumin — the most abundant plasma protein — to pass through.
"We wanted to see if maybe there were some things that we were looking at through this type of sample prep, that others have not," Matthew Kuruc, VP of business development at BSG, said in an interview.
As part of that study, BSG also began investigating the possibility of common proteomic patterns across cancers of diverse tissue types.
"There weren't a whole lot of studies on this," Kuruc said.
The screen did turn up several proteins that fit this pan-cancer pattern, most of which related to inflammation and more specifically, to innate immunity.
In particular, the BSG screen turned up truncated — and therefore inactive — serpins, a family of protease inhibitors whose dysregulation associates with cancer progression.
Although whether serpin dysregulation leads to inflammation and immune dysregulation or the other way round remains a chicken-and-egg problem, the data led BSG to develop a hypothesis that certain innate immunity pathways — specifically coagulation, complement, and neutrophil recruitment — drive stromal conditioning within the tumor microenvironment.
BSG reasoned that the failures of other common protein screening methods such as ELISA to discriminate between active and inactive isoforms could obscure relationships between serpin abundance and disease phenotype.
Measuring serpin abundance without regard to their activity status, Kuruc says, "is misleading because you're not measuring function."
BSG now has a patent pending on a method for discriminating between active and inactive serpin proteoforms, and is looking for partners to help explore commercial product and biomarker opportunities related to them.
Also in 2017, Wilma Mesker, an assistant professor in the Department of Surgery at LUMC, and her colleagues published a study, identifying the tumor-stroma ratio — the proportion of tumor cells relative to the surrounding stroma — as an independent risk factor for poor overall, metastasis-free, and recurrence-free survival in primarily operable breast cancer.
Mesker and her colleagues found that tumors from patients with worse prognoses contained significantly more stroma.
"When I read about the Stroma Liquid Biopsy from BSG," she said via email, "I immediately contacted them, resulting in a nice collaboration to evaluate stroma formation in tissue and biomarkers in blood."
BSG and LUMC inked a collaboration agreement in 2019, under which BSG would define and quantify the SLB proteins in LUMC-collected sera in order to correlate the LUMC-developed tumor-stroma ratio scoring methods.
In the current study, Mesker, Kuruc, and colleagues analyzed transcriptional abundances of the 13 SLB genes from two public repositories: The Cancer Genome Atlas and the Gene Expression Omnibus.
By performing gene set enrichment analysis on a discovery cohort of 359 colon adenocarcinoma patient and a validation cohort of 229 such patients, the team found that SLB transcripts were significantly enriched in tumors with high histologic stromal content. From single-cell sequencing data, they also identified stromal-specific and epithelial-specific SLB gene expression patterns, resulting in a prognostic stromal-epithelial gene expression signature ratio.
Among the SLB panel of genes, the researchers found that TIMP1, PF4, and SERPINA1 associated with patient survival and that all three participate in platelet degranulation, a characteristic response to coagulation, which itself associates with various cancer types.
"I was quite excited about this observation since it may have serious implications for our serum SLB proteomic profiles," Mesker said. "For instance, dysregulated coagulation proteins in serum, as measured by the SLB panel, may be indicative of high intra-tumoral stromal content and, hence, poor patient prognosis."
These results contributed to a prognostic risk model, in which tumors with over 50 percent stroma content (stroma-high tumors) carry a significantly worse prognosis than stroma-low tumors.
If validated in further studies, the LUMC and BSG researchers hope that the SLB can help guide therapies that aim to modulate the tumor microenvironment.
"The variety of potential new therapeutic strategies that can be monitored by the SLB panel is numerous and varied," Kuruc said.
Several anticoagulants have been clinically tested as potential cancer treatments, for instance, especially for early-stage cancers. Reactive center peptides from serpins have also shown potential immunomodulatory functions, suggesting that selectively tuning their activity could help modulate stroma formation.
"In general, modulation of stromal conditioning might improve immuno-oncology treatments, for example turning 'cold' tumors to 'hot,'" Kuruc said. "As all these innate pathways are interconnected, modulation of one might help unwind the rest."
With the results of the current work providing a theoretical framework for using proteomic signatures to potentially report on tumor-stroma content in a liquid biopsy, the collaborators have begun a study to directly test the SLB on patient sera.
While final results have yet to be published, Mesker mentioned having observed increased serum platelet counts in patients with stroma-rich tumors, lending support to the idea that serum content may reflect tumor microenvironment characteristics.
"We are currently fine-tuning the test for the analysis of patient samples for clinical use," Mesker said. "A cohort of patients from the Dutch national screening program for colorectal cancer will be measured, including precancerous lesions, such as high-risk polyps."
Although the group hopes to bring SLB to clinical trials and eventual commercialization, they must first await the results of ongoing preclinical studies before making firm plans.
"It's too early to know what the best commercial path is," Kuruc said, adding that while BSG and LUMC currently lack any commercial partners, they welcome all inquiries aimed at helping to commercialize SLB.
Meanwhile, BSG has been exploring the use of the SLB beyond cancers. Last year, the company presented a poster at American Society for Mass Spectrometry conference, in which they used the SLB to characterize serum profiles of hospitalized COVID-19 patients. Here again, they found changes in the ratios of active and inactive serpins correlating with disease status, providing both a useful test of the SLB and new data on how SARS-CoV-2 infection plays out.