Immune checkpoint molecules regulate immunosuppressive pathways to maintain self-tolerance and prevent excessive immune activation, but their overexpression can inhibit effector immune cells, aiding cancer growth. Targeting these molecules with immune checkpoint inhibitors (ICIs) has shown promise in cancer therapy, especially against non-small-cell lung cancer (NSCLC). The PD-1 receptor, expressed on various immune cells and cancer cells, promotes immune balance under normal conditions but contributes to T cell exhaustion in tumors. ICIs targeting the PD-1/PD-L1 pathway have improved survival in NSCLC patients, though responses vary, with 60-70% of patients progressing within six months. Tumor-infiltrating lymphocytes (TILs) have emerged as predictive biomarkers for ICI response, yet antibodies used in ICIs have limitations in size, clearance, and tumor penetration. To address these issues, we developed small protein variants (MBA series) using ribosome display, which show high affinity and specificity to PD-1. These MBA proteins can monitor PD-1+ cell populations in tumor biopsies and image tissues with PD-1+ cells, offering a potential diagnostic tool for NSCLC management.

Materials and Methods

Myomedin Combinatorial Library Design

The Myomedin β-sheet combinatorial library was based on human myomesin-1 domain 10. In silico mutability screening identified residues for randomization using FoldX, assigning a mutability score to each residue. Scores were calculated for stable myomesin-1 domain 10 structures (PDB IDs 3rbs, 6t3o).

Construction and Assembly of Myomedin β-Sheet Library

A series of PCR steps, using Q5® High-Fidelity DNA Polymerase (NEB) and specific primers, constructed the Myomedin β-sheet library. PCR products were gel-separated, and desired lengths were extracted. Three PCR rounds were used: internal non-randomized library part, attaching randomized parts, and completing the Myomedin scaffold. Ribosomal display elements were added in two more PCR steps.

Ribosome Display

Three rounds of ribosome display selection were performed using human PD-1 (R&D Systems) coated on MaxiSorp 96-well plates. Washing was done with increasing stringency in TBS-T. Post-selection, cDNA was cloned into pET28b vectors, inserted into E. coli XL1-blue, sequenced, and screened for hPD-1 binding using ELISA.

In Silico Modeling by Docking

MBA variants were modeled with MODELLER 9v14 based on myomesin-1 domain 10 structure (PDB ID 3rbs). The extracellular domain of hPD-1 and hPD-1/hPD-L1 complex models were built using AlphaFold. Docking was performed with ClusPro, and results were analyzed using PyMOL.

Bioconjugation of MBA Binders

MBA proteins were conjugated with p-NCS-Bz-DFO. After dialyzing into PBS, the pH was adjusted, and p-NCS-Bz-DFO was added. The mixture was incubated, unreacted deferoxamine removed, and buffer exchanged using a Zeba spin desalting column.

68Gallium-Labeling of MBA Proteins

DFO-conjugated binders were radiolabeled with 68Gallium. The mixture was incubated, and radiochemical purity was assessed by iTLC chromatography using sodium citrate and sodium acetate-methanol mobile phases.

In Vivo and Ex Vivo Analysis of Biodistribution in Mouse Tissues

Animal experiments were approved and conducted per Czech regulations. Mice were injected with labeled binders and sacrificed at specified times. Tissue samples were collected, weighed, and radioactivity measured. PET/CT imaging was conducted with dynamic and static scans post-injection.

Mouse Infection Model

A mouse model of acute myositis was established by injecting E. coli into the left hind leg. After infection development, mice were injected with 68Ga-labeled binders and imaged by PET/CT.

Immunohistochemistry Staining

Tissues were frozen, sectioned, fixed, and blocked. MBA414 or anti-PD-1 antibody-FITC was applied overnight, followed by anti-V5 Alexa Fluor 647 antibody. Mounting medium with DAPI was used for microscopy. Co-localization was quantified using ImageJ software.

Results from nanoScan® PET/CT

Biodistribution of Radiolabeled MBA Proteins and In Vivo PET/CT Imaging

Myomedin variants MBA066, MBA197, MBA414, and control MyoWT were selected for biodistribution analysis in mice. Prior to radiolabeling, these proteins were conjugated with Deferoxamine (DFO) using p-SCN-Bz-DFO to target lysine residues. They were then labeled with 68Gallium and dialyzed into PBS. The radiochemical purity of labeled proteins reached around 95%, suitable for clinical diagnostics, and their stability in human serum was confirmed to be 89.6-98.2% over 2 hours.

In vivo biodistribution was tested in BALB/c mice using PET/CT imaging up to 90 minutes post-injection. 68Ga-MyoWT accumulated rapidly in the kidneys and was excreted via the bladder, serving as a negative control. MBA066 showed rapid heart accumulation within the first 25 minutes and significant liver presence throughout. MBA197 and MBA414 accumulated substantially in the liver with slow clearance through the kidneys, indicating specific interactions with PD-1+ cell populations, primarily in the liver.

Fig.4. In vivo imaging and ex vivo analysis of ⁶⁸Ga-labeled MBA distribution in Balb/c mice. A ⁶⁸Ga-MyoWT, B ⁶⁸Ga-MBA066, C ⁶⁸Ga-MBA197, D ⁶⁸Ga-MBA414. Mice were retro-orbitally injected with particular MBA binders and immediately scanned by a PET/CT scanner. PET imaging was performed in the form of eighteen consecutive 5-min PET scans. E ⁶⁸Ga-myoWT, F ⁶⁸Ga-MBA066, G ⁶⁸Ga-MBA197, H ⁶⁸Ga-MBA414. Mice were retro-orbitally injected with ⁶⁸Ga-binders (n = 3 mice per group). Blood, spleen, pancreas, stomach, intestines, kidneys, liver, heart, lungs, muscle, and bone were collected at 30 and 90 min after injection. The tissue samples obtained were weighed, and radioactivity was measured in a γ-counter. The biodistribution data were calculated as the percentage of injected dose per gram of tissue (% ID/g) In vivo imaging and ex vivo analysis of 68Ga-labeled MBA distribution in Balb/c mice. A 68Ga-MyoWT, B 68Ga-MBA066, C 68Ga-MBA197, D 68Ga-MBA414. Mice were retro-orbitally injected with particular MBA binders and immediately scanned by a PET/CT scanner. PET imaging was performed in the form of eighteen consecutive 5-min PET scans. E 68Ga-myoWT, F 68Ga-MBA066, G 68Ga-MBA197, H 68Ga-MBA414. Mice were retro-orbitally injected with 68Ga-binders (n = 3 mice per group). Blood, spleen, pancreas, stomach, intestines, kidneys, liver, heart, lungs, muscle, and bone were collected at 30 and 90 min after injection. The tissue samples obtained were weighed, and radioactivity was measured in a γ-counter. The biodistribution data were calculated as the percentage of injected dose per gram of tissue (% ID/g)

To further examine the biodistribution of PD-1-binding Myomedins, we conducted post-mortem ex vivo analysis of 68Ga-Myomedins in multiple mouse organs. Quantification of radiolabeled MBA066, MBA197, MBA414, and MyoWT was performed 30 and 90 minutes post-administration in blood and ten organs. MyoWT predominantly accumulated in the kidneys, with weak signals in the liver, spleen, and blood. In contrast, MBA197 and MBA414 showed significant concentration in the liver (6-7 times), spleen (4-7 times), and lungs (8-10 times), but minimal accumulation in other organs.

To test if MBA197 and MBA414 specifically target PD-1-expressing cells, we used a mouse muscle infection model with intramuscular E. coli injection. 68Ga-MBA066, primarily binding to liver cells but clearing faster via kidneys, showed distinct signal concentration at the infection site 45 minutes post-injection compared to 68Ga-MyoWT. This supports the hypothesis that 68Ga-MBA066 specifically targets PD-1+ lymphocytes, which accumulate at infection sites.

In summary, the use of radiolabeled Myomedins for PET/CT imaging shows great promise for non-invasive monitoring of PD-1+ cell populations in vivo. The distinct biodistribution patterns observed for different Myomedin variants highlight their potential for targeted imaging and diagnostics, offering a valuable alternative to traditional antibody-based approaches. Further studies are warranted to fully explore their clinical applicability and effectiveness in predicting responses to immunotherapy.

Full article on ncbi.nih.gov