Checkpoint Protein Analysis and Characterization

The complexity of emerging biopharmaceuticals, their involvement in labyrinthine cellular and molecular pathways, and the imperative for exactness in precision medicine underscore the importance of good analytics. This article could therefore have been distilled down to one idea: liquid chromatography/mass spectrometry (LC-MS), the standard method in quantitative proteomics, and whatever else—cell counting, 'omics, etc.—as needed.

While standard, accessible LC-MS methods have supplanted (but not replaced) biochemical assays in proteomics, including in the study of checkpoint inhibitor proteins and the molecules with which they interact, clinical applications demand much higher robustness and reliability than do basic investigation.

The major players

Immune checkpoints regulate interactions occurring during an immune response, between immune system T cells and antigens. These events occur either through co-stimulatory or co-inhibitory processes. Co-stimulatory proteins include CD28, ICOS, and CD137, each with a specific role.

Ligation of CD28, coupled with T-cell receptor (TCR) activation, initiates the Immune response. Both are necessary because while TCR activation discriminates against "self" and "non-self" quite well, it also requires a strong CD28 signal to differentiate between pathogenic and benign antigens.

ICOS (inducible T-cell co-stimulator) is a surface marker on activated T cells that binds to a specific ligand on antigen-presenting cells. Binding of anti-ICOS antibodies on tumor-infiltrating T cells stimulates ICOS-positive T-cell expansion, and enhances both T-lymphocyte survival and its tumor cell responses. Anti-ICOS monoclonal antibodies such as JTX-2011 and GSK3359609, are development-stage drugs while pembrolizumab (Keytruda) was approved in 2014 for various solid and blood cancers and continues to expand its label.

The checkpoint regulator CD137 (also known as 4-1BB) is expressed in both natural killer (NK) cells and T cells, so it induces both innate and adaptive immunity. Upon exposure of immune cells to a tumor antigen, CD137 stimulates T and NK cells to expand and collectively ramp up their antitumor activity. Because of its role in inflammation and its expression in blood vessels, CD137 has been implicated in development of atherosclerotic plaque, a precursor to heart attack and stroke.

By contrast, co-inhibitory immune checkpoint players PD-1, CTLA-4, and VISTA inhibit the immune response by protecting healthy tissues from targeting and destruction. This is the mechanism hijacked by cancer cells, which dysregulate inhibitory processes in ways that allow cancer cells, like normal "foreign" cells or molecules, to evade checkpoints.

VISTA, a recently discovered checkpoint pathway found exclusively in blood lineage cells, directly suppresses T-cell activation.

By far the most-studied immune checkpoint proteins, and the targets of several breakthrough drugs, are PD-1 (programmed cell death 1) and CTLA-4 (cytotoxic T-lymphocyte-associated antigen 4).

PD-1 incorporates two ligands, PD-L1 and PD-L2. PD-L1 dominates, and is expressed in hematopoietic cells including T cells, B cells, dendritic cells, macrophages, and mast cells, and in numerous other cells including endothelial and epithelial and in many tumors. Its main role is to inhibit anti-tumor immune responses by T cells.

CTLA-4 is homologous, and shares two ligands with CD28 but carries out the opposite function. Where CD28 mediates T-cell co-stimulation, interactions with CTLA-4 inhibit T-cell responses.

The dominance of PD-1 and CTLA-4 notwithstanding, these are by no means the only checkpoint proteins, or even the only significant ones. Many new drugs under evaluation target several immune checkpoint proteins, including LAG3, TIM3 (HAVCR2), CD40, and GITR (TNFRSF18). To complicate matters, several clinical trials are studying pembrolizumab combined with antibodies against LAG3, RORC, CTLA-4, and with small molecule inhibitors of JAK1, EGFR, MEK1/2, with vaccines, and even patient-specific tumor antigens.

Characterization of ICIs

Methods based on liquid chromatography (LC), especially with mass spectrometry, currently dominate instrumental proteomics, including the quantitation of individual endogenous checkpoint proteins and administered checkpoint inhibitor monoclonal antibodies. These methods are published and accessible. But as more checkpoint inhibitor drugs enter clinical trials and are eventually approved, the scope of analysis requirements greatly expands. Methods developed to test homogeneity during quality control of a checkpoint inhibitor protein in buffer will require sample preparation when performed on blood, and would likely involve modifying elution conditions when more than one protein species was under investigation. Clinical protocols may require monitoring secondary players, such as chemical signals, proteomic byproducts, or metabolites, as well as populations of up- or downregulated immune system cells and any molecular or cellular species that could serve as a biomarker for treatment efficacy.

More information is always better than less for drug developers and regulators, but analytics for checkpoint inhibitor treatments serve a higher purpose than simply to optimize yields or assure quality. What we analyze, and the method's reproducibly and robustness, affect development and approval strategies, and eventually govern the apparent efficacy of the treatment.

European researchers raised these issues by asking if current CPI inhibitor studies could benefit from new methodologies:

"…the choice of the most accurate statistical methods, endpoints and clinical trial designs to estimate the benefit of ICI remains an unsolved methodological issue. Considering the unconventional patterns of response or progression … observed with ICI, the application in clinical trials of novel response assessment tools … is an unmet clinical need."

While the authors are referring to measures of efficacy, their question could just as easily have been directed toward analytical scientists instead of physicians. The choice of statistics, endpoints, and trial take into account, after all, the developer's analytic capabilities.

For example, LC-MS methods for quantifying levels of the checkpoint inhibitor nivolumab in laboratory assays are readily adapted to plasma samples from patients undergoing nivolumab treatment for non-small cell lung cancer. One study showed reasonable (to within 15%) agreement with the non-clinical method over a concentration range of 5–200 mcg/mL.

However, in a clinical setting, researchers may be looking for more than just checkpoint inhibitor levels, or for limits of detection below those which standard methods reliably provide. The flip side of nivolumab efficacy is the drug's side effects, which target multiple organ systems. Predicting which patients will experience these adverse events may require a time-based comparison, i.e., pharmacokinetic study, for which "total nivolumab" might not tell the entire story.

Obtaining precise numbers, including concentration changes of clinical significance, requires a more nuanced approach involving internal standards, concentration curves, carryover, and response ratio calculations to correct for sample and method variability. For these studies, either the intact protein or an easily identified surrogate or proteolysis product is required. Fortunately, the nivolumab Fab region fits the bill from the surrogate/analysis perspective and is also accessible via standard proteolysis methods.

Japanese researchers have developed an LC-MS method based on nano-surface and molecular-orientation limited Fab-sensitive proteolysis. This approach, which targets the signature peptide ASGITFSNSGMHWVR on the protein's complimentarity-determining region, turns a research-oriented analysis into a validated assay that is " applicable as a method for clinical PK and PK-guided cancer therapy for Nivolumab."

Conclusion

This brief introduction to checkpoint inhibitor characterization was not designed to advise seasoned analysts on the art of LC-MS-based proteomics. The purpose was instead to point out that:

• Checkpoint inhibition is complex, probably involving several pathways and yet-to-be-discovered targets
• While individual checkpoint proteins and inhibitors are remarkably diverse in their activity, they are all comprised of amino acids and are therefore susceptible to conventional proteomic LC-MS
• Frontiers in checkpoint inhibitor analysis will involve adapting and optimizing today's analytic paradigm, proteomic LC-MS, but realizing actionable diagnostics will require extensive validation and tweaking
• Opportunities exist outside the proteomics paradigm in genomics, but especially in metabolomics, for discovery of surrogate markers for both checkpoint proteins/checkpoint protein inhibitors, and for treatment efficacy