Edgar Gonzalez-Kozlova, PhD, Emphasizes the Potential for Circulating Soluble Proteins to Predict Response to Immunotherapy in Lung Cancer

Blood circulating soluble proteins may be an effective and accessible biomarker for predicting response to immunotherapy among patients with squamous cell lung cancer.

Results from serial blood specimens in certain patients treated on the phase 3 SWOG Lung-MAP S1400I trial (NCT02154490) indicated that blood circulating soluble proteins could be a promising predictor of outcomes among patients with squamous cell lung cancer, according to Edgar Gonzalez-Kozlova, PhD.

Data from the substudy, which set out to assess how changes in serum analyte levels may be associated with outcomes following treatment with nivolumab (Opdivo) plus or minus ipilimumab (Yervoy) in squamous cell lung cancer, indicated that serum proteins PCDC1, CXCL9, and CXCL10 had increased from baseline to weeks 3, 7, 9, and upon progression. Moreover, CCL19 increased at weeks 3 and 7 but not week 9 or at progression. Changes in CXCL13 were most significant from baseline to progressive disease among those treated with the doublet regimen vs nivolumab alone. Those with an objective response appeared to have higher IL4/LAMP3 and lower IL6/8 at baseline and week 3 vs those with no response.

“Take care not to forget proteins because they are important,” Gonzalez-Kozlova cautioned. “Unfortunately, right now, a lot of studies focus on a single point, so it’s easy to miss other assays or technologies that look at other [markers]. I would encourage everybody to try to use this technology when they are trying to combine it with other assays. For example, one of our trials here has many acids; we’re looking at proteomics, transcriptomics, cytometry, and even histology. The idea is always to, in the future, go to our systems biology where we can see exactly what’s happening in all aspects of the disease to better understand it.”

In an interview with CancerNetwork® during the 2022 American Society of Clinical Oncology Annual Meeting, Gonzalez-Kozlova, an assistant professor at the Icahn School of Medicine, Mount Sinai, highlighted potential proteins that may be accessible predictors of response and the direction biomarker research may be heading.

CancerNetwork®: What was the rationale for assessing changes in serum analyte levels to predict clinical outcomes with nivolumab and ipilimumab in lung cancer?

Gonzalez-Kozlova: For a long time, proteomics have been very expensive and complicated to [use]. There has been a lot of progress in the last few years. At Mount Sinai, we have 1 of the few machines that performs this kind of assay. [The technology] is really interesting because we basically still use antibodies to detect proteins, but it’s at very low quantities and very small volumes. These antibodies are attached to the ctDNA that allow us to run a PCR [polymerase chain reaction test]. Basically, we can amplify the signal a few hundred times for these small analytes. That’s how we detect the circulating proteins from blood. It’s really exciting because it’s very sensitive, and because it’s a cheap, we can have a lot of control.

This allows us to detect up to 92 proteins for a specific panel we want, or we can also design our own and keep track of these patients. That was the rationale of why we want to see these proteins. A lot of these changes happen at the protein level and sometimes when we are detecting them at the RNA level, [we are] a bit too early or too late. Proteins give us a good snapshot of what’s happening at that precise moment in time for these patients. Fortunately, this clinical trial had longitudinal follow up, meaning that the patients had many visits with the physicians, including one before starting the treatment, so we [were able to include] baseline [findings]. This was very exciting because we can now know exactly what is changing before the patients receive treatment. Even more exciting is that this trial has 2 arms: 1 that receives only nivolumab and we also have a second arm that has both ipilimumab and nivolumab. It’s a PD-1 plus CTLA-4 checkpoint blockade [regimen]. These components [cause breaks in] the immune system, and by blocking them using these monoclonal antibodies, we can basically stop the breaks and allow the immune system to keep doing its work of cleaning up the cancer.

What were the key findings of this study?

After we followed up on what was happening with these patients, we saw that there was a set of proteins [with increasing levels and others that were decreasing] after treatment. Patients who ended up having a good response had a lower number of inflammation [markers], which is interesting because usually inflammation helps to attract immune cells to clear the cancer. [Responders] had markers for inflammation like IL6 or [CXCL13]. However, the [volume of these markers in responders] was always a little bit lower vs patients who did not have a response. We have also been seeing the involvement of very interesting proteins, as well. Some of them have been reported in literature, but it was previously unclear if they [were predictive of response] or not. One of them is actually another molecule that we think helps activate T cells. It’s a sugar receptor for metabolism of glucose called ICOS. Basically, it helps with glucose uptake for this disease. When it’s upregulated, usually the T cells are a bit more dormant and less reactive. But in this case, patients who responded [had higher levels] than patients who did not respond, and all of the inflammation markers were also lower in these patients. This means that inflammation is good but a little too much is bad. We learned this lesson from the COVID-19 pandemic recently. There were many other markers that were associated with T-cell response and T-cell migration that were always present, but these were lower than in patients who did not respond. The main finding of this was that even in cancer, a little too much inflammation is bad.

Where will biomarker research be focused in the future?

Markers that seem to be promising are CXCL13, some matrix metalloproteinases like MMP12, and CSF1, which are very interesting. Sometimes you also see LAMP, which is a dendritic cell marker that is exciting, as well. [With regard to] future research, the study is funded by Sacha Gnjati, PhD, of Icahn School of Medicine. He is the leading primary investigator and manages most of the grants and funding on this trial while I handle the research work. We have many trials using these grants; this is a U24 network grant. We work very closely with CMAX, which is a cancer immunology network. We have over 30 trials, this being one of them. Basically, we’re trying to look at what happens overall, across all cancers, and across all trials to see if there is a strong relationship with therapy or response.

Future plans are to complete other trials, and we currently have completed 5 and have run assays for about 20. There are a few more to run and to analyze, but once we have all of the data, we will be [reporting] stronger claims about these markers. This can potentially help because when we look at markers before treatment, [collecting these data is a] very good thing for the patients even if we cannot separate who will respond and who will not. Just [being able to] stratify the patients [based on] risk is already a plus. That’s exactly the goal, to help these patients. And this study will help point us in the right direction of what to look for in the clinic.

Reference

Gonzalez-Kozlova E, Huang HH, Redman MW, et al. Dynamic changes in serum analyte levels associated with clinical outcome in squamous cell lung cancer trial SWOG Lung-MAP S1400I of nivolumab ± ipilimumab. J Clin Oncol. 2022;40(suppl 16):9044. doi:10.1200/JCO.2022.40.16_suppl.9044