The ProtParam tool (https://web.expasy.org/protparam) [5] was employed to assess the physicochemical properties of unique IGs, especially their molecular stability. All 46 IGs identified through Venn diagram analysis were subjected to ProtParam screening to refine the selection. This systematic approach enabled the identification of the most stable IGs, highlighting promising candidates for future therapeutic development and biomarker discovery.

The three-dimensional structure of K-Ras was retrieved from the AlphaFold Protein Structure Database (EMBL-EBI, UK). Protein entries were filtered by organism (Homo sapiens) and annotation status to ensure the use of a validated, manually reviewed structure (Swiss-Prot). Based on these criteria, the final structure selected for downstream analyses was GTPase K-Ras (AlphaFold ID: AF-P01116-2-F1-v6). The structure was accessed from the AlphaFold database (https://alphafold.ebi.ac.uk/entry/AF-P01116-2-F1).

To evaluate the therapeutic potential of the unique IGs, blind molecular docking was performed using CB-Dock2 (https://cadd.labshare.cn/cb-dock2/php/blinddock.php) [6]. Structural data files (SDF) for six FDA-approved CRC drugs—Irinotecan (IRN), oxaliplatin (OXA), 5-fluorouracil (5-FU), lonsurf (trifluridine-tipiracil) (FTD-TPI), capecitabine (CAP), and leucovorin (LEU)—were downloaded from PubChem and used as ligands for docking. CB-Dock2 facilitated the structure-based evaluation of IG-drug interactions, while AutoDock Vina refined the binding conformations, offering insights into the strength and stability of IG-drug complexes.

Epitope prediction for both IG heavy variable 3-64 (IGHV3-64) and K-Ras was performed using the Immune Epitope Database (IEDB; https://www.iedb.org/) [7]. For modeling epitope-epitope interactions, we used CABS-dock (https://biocomp.chem.uw.edu.pl/CABSdock/) [8] under standard simulation settings, allowing flexible peptide-protein docking to simulate native-like interactions and assess the immunological and structural compatibility between IGHV3-64-derived epitope peptides and K-Ras neoantigenic epitopes. As four IGHV3-64 epitopes were identified, each was individually screened in separate docking simulations.

Molecular simulation setup:

K-Ras protein sequence(s):

MTEYKLVVVGAGGVGKSALTIQLIQNHFVDEYDPTIEDSYRKQVVIDGETCLLDILDTAGQEEYSAMRDQYMRTGEGFLCVFAINNTKSFEDIHHYREQIKRVKDSEDVPMVLVGNKCDLPSRTVDTKQAQDLARSYGIPFIETSAKTRQRVEDAFYTLVREIRQYRLKKISKEEKTPGCVKIKKCIIM

IGHV3-64 peptide sequence: EVQLVESGEGLVQPGG

As the objective is to find epitope-epitope pairing between K-Ras and IGHV3-64-derived epitope candidates (short peptides), intending to reveal naturally compatible or selective interactions that could be harnessed for precise delivery of polyclonal drug conjugates (pPDCs). We investigated the binding interactions between the K-Ras protein and the IGHV3-64 epitope, Monte Carlo (MC) simulations were performed. A total of 50 MC cycles were conducted to assess the structural stability and conformational dynamics of the complex.

To determine whether the unique IGs possess cell-penetrating capabilities, we analyzed their corresponding full-length protein sequences—sourced from UniProt—using the CellPPD server (http://crdd.osdd.net/raghava/cellppd/) [9]. This analysis identified putative CPP regions, indicating peptides with potential for enhanced intracellular delivery.

The identification of IGs specific to different CRC stages requires high-resolution analysis of human plasma samples. IGs are membrane-bound or secreted glycoproteins produced by B lymphocytes. During the recognition phase of humoral immunity, membrane-bound IGs function as antigen receptors. Upon antigen binding, they initiate clonal expansion and differentiation of B cells into plasma cells, which subsequently secrete IGs. These secreted IGs mediate the effector phase of humoral immunity by neutralizing and eliminating bound antigens. The variable (V) region of the IG heavy chain plays a key role in antigen recognition. Motivated by this functional specificity, we initiated a discovery-phase proteomics study to identify circulating, stage-specific IGs. A total of 12 pre-treatment plasma samples, representing clinical tumor, nodes, and metastasis (TNM) stages I–IV, were analyzed using high-resolution ESI-nanoLC-MS/MS. An overview of the experimental workflow is presented in Figure 1.

Mass spectrometry analysis identified a total of 138 IGs across the four CRC stages, with 64, 25, 25, and 24 IGs detected in stages I, II, III, and IV, respectively. To clarify the distribution of IGs across disease stages, Venn diagram analysis was performed to distinguish stage-specific IGs—defined as IGs uniquely detected in only one CRC stage—from overlapping IGs shared between two or more stages (Figure 2).

Physicochemical stability analysis assesses a protein’s physical and chemical properties to evaluate its suitability for therapeutic or diagnostic use—key for ensuring efficacy, safety, and manufacturability. Parameters such as amino acid (AA) composition, molecular weight, isoelectric point (pI), and hydrophobicity were evaluated using ProtParam. All 46 stage-specific IGs were analyzed, and 5 were identified as stable—3 from early-stage CRC (ESCRC; 2 from stage I and 1 from stage II) and 2 from advanced-stage CRC (1 each from stages III and IV), as summarized in Table 2. These 5 IGs met the defined stability criteria, highlighting their potential for drug conjugation and further clinical development.

The 5 prioritized IGs underwent extensive molecular docking analysis against 6 FDA-approved CRC chemotherapies: IRN, OXA, 5-FU, FTD-TPI, CAP, and LEU. This analysis aimed to determine potential drug-specific binding affinities and ultimately identify promising pPDCs. These FDA-approved drugs form the backbone of standard chemotherapy regimens and are widely used across many countries, earning their place as core chemotherapy agents in the WHO Essential Medicines List (WHO-EML). They play critical roles in treatment protocols—either as monotherapy or in combination—depending on cancer stage, patient-specific factors, and therapeutic intent (curative vs. palliative) (Table 3).

Based on the top binding scores, IGHV3-64 was selected for further analysis due to its strong binding affinity not only to IRN but also to several other CRC chemotherapeutic agents. This finding prompted us to investigate whether IGHV3-64 shares any molecular epitopes with known CRC oncogenes, and whether it contains any CPP sequences—two critical factors that could support its development as a pPDC candidate. To explore potential pathways for pPDC design and enable downstream functional analysis, we focused on K-Ras, one of the most frequently mutated oncogenes in CRC (Figure 4).

K-Ras, a pivotal oncogene implicated in a range of malignancies such as colorectal, pancreatic, and non-small cell lung cancer, continues to be a notoriously difficult protein to drug due to its high structural flexibility and absence of deep, druggable pockets. Despite the approval of covalent G12C inhibitors, many K-Ras mutations remain untargetable, necessitating new therapeutic strategies [14–16]. Here, we explore a novel approach based on epitope-epitope molecular recognition, specifically between short peptide epitopes derived from the IGHV3-64 variable domain and functional epitope regions of K-Ras. These interactions are investigated as a platform to guide the rational design of peptide drug conjugates (PDCs) for intracellular precision delivery.

Following epitope pairing analysis, we investigated the presence of CPP sequences within IGHV3-64 and K-Ras to evaluate their theoretical potential for intracellular delivery. Identification of CPPs within CRC-associated proteins provides a bioinformatic strategy to assess whether endogenous protein fragments possess physicochemical properties compatible with membrane translocation.

CPP prediction was performed using the CellPPD server, which applies a support vector machine (SVM)-based classification model. In accordance with the tool’s recommendations, peptides with SVM scores greater than 0.0 were considered CPP-positive, while negative scores indicated non-CPP sequences. This threshold was applied uniformly across all analyzed peptides and is now explicitly reported to improve methodological transparency. CPP profiling revealed that both IGHV3-64 and K-Ras harbor multiple peptide segments predicted to possess cell-penetrating properties (Table 6). These peptides demonstrated moderate-to-high SVM scores, balanced hydrophobic-hydrophilic profiles, and compact structural features—attributes commonly associated with membrane interaction and intracellular access. Within IGHV3-64, two CPP-positive peptides were identified: MHWVRQAPGK (AA 54–63) and KGRFTISRDN (AA 85–94). These peptides exhibit relatively low molecular weights (~1.2 kDa), basic pIs (~pI 11), and modest amphipathicity, suggesting theoretical compatibility with membrane interaction and endosomal escape under acidic intracellular conditions. Their physicochemical profiles also indicate solubility and chemical accessibility suitable for conjugation with small-molecule payloads such as IRN or OXA. K-Ras contained multiple overlapping CPP-positive regions, particularly within its C-terminal sequence. The peptide RQYRLKKISK (AA 164–173) exhibited the highest SVM score (0.68), along with strong amphipathicity (1.72), a high net positive charge (+5), and a basic pI (11.17). These characteristics are consistent with electrostatic interactions with negatively charged cancer cell membranes. Additional CPP-positive peptides, including REIRQYRLKK and VREIRQYRLK, displayed similar cationic and amphipathic profiles, further supporting their predicted membrane-interactive capacity.

From a pPDC design perspective, the presence of CPP-positive regions in both IGHV3-64 and K-Ras supports the theoretical feasibility of constructing conjugates capable of intracellular access without reliance on synthetic delivery vectors. However, it is important to emphasize that CPP prediction represents a computational feasibility assessment, not experimental validation. Comparative cytotoxicity assays (e.g., free IRN vs. IGHV3-64-conjugated formulations), cellular uptake studies, and intracellular localization analyses will be essential in future work to confirm functional delivery.

This study introduces a novel multi-omics framework for identifying circulating, stage-specific IGs in CRC, with the ultimate aim of enabling precision-guided pPDC design. High-resolution ESI-nanoLC-MS/MS plasma proteomics across TNM stages I–IV identified 138 IGs, of which 46 were uniquely stage-specific. Through physicochemical filtering based on molecular weight, pI, and stability indices, five stable IGs representative of ESCRC and advanced-stage CRC (ASCRC) were prioritized for downstream analyses. The stage-specific enrichment of circulating IGs is unlikely to be stochastic and instead reflects dynamic immune remodeling within the CRC TME during disease progression. CRC evolution is accompanied by marked shifts in immune cell composition, cytokine gradients, and antigenic burden. While ESCRC is often associated with effective immune surveillance, ASCRC is characterized by a chronically antigen-exposed yet immunosuppressive TME, marked by sustained inflammation, immune exhaustion, and altered lymphocyte functionality.

Within this evolving TME, B cells and humoral immune responses play increasingly recognized and context-dependent roles. ASCRC has been associated with expanded B cell infiltration, tertiary lymphoid structure formation, and clonal expansion of antibody-producing plasma cells. Persistent exposure to tumor-associated antigens and neoantigens—particularly mutant K-Ras—can drive selective B cell activation and biased IGHV gene usage, resulting in repertoire skewing toward specific IGHV families. Accordingly, the detection of stage-specific circulating IGs, including IGHV3-64 in advanced CRC, likely reflects antigen-driven humoral immune adaptation rather than a passive by-product of tumor burden.

From a translational perspective, this stage-associated IG pattern also suggests potential diagnostic and stratification value. Although receiver operating characteristic (ROC) and area-under-the-curve (AUC) analyses were not performed due to limited cohort size, the consistent enrichment of IGHV3-64 in ASCRC supports its candidacy for future validation as a circulating biomarker to distinguish early vs. advanced disease in larger, independent cohorts.

The concept of employing endogenous IGs as precision-targeting agents aligns with the established clinical success of ADCs in solid tumors, including HER2-positive malignancies [17]. Our findings extend this paradigm by identifying tumor stage-associated circulating IGs shaped by the CRC immune microenvironment and proposing their repurposing as endogenous scaffolds for pPDC development—potentially offering a personalized and immunologically compatible alternative to monoclonal ADCs, particularly in heterogeneous and immune-evasive advanced CRC.

Molecular docking analysis revealed that several prioritized IGs—particularly IGHV3-64, IGHV3-23, and IGLV9-49—exhibited selective binding affinity toward IRN. Among these, IGHV3-64 demonstrated the strongest predicted interaction (ΔG = −10.0 kcal/mol), highlighting its feasibility as a drug carrier framework. This observation is consistent with prior studies demonstrating that antibody-based delivery systems can enhance the bioavailability and therapeutic efficacy of chemotherapeutic agents in CRC xenograft models [18]. The preferential association of IGHV3-64 with IRN, relative to other shortlisted IGs, likely reflects distinct structural and physicochemical features of its variable region rather than a generic property of IGs. Specifically, favorable charge distribution, hydrophobic pocket architecture, and conformational flexibility within the complementarity-determining regions (CDRs) may facilitate stable non-covalent interactions with small-molecule chemotherapeutics. Such features may arise through antigen-driven clonal selection in advanced CRC, further supporting the functional relevance of IGHV repertoire skewing.

Importantly, this differential binding does not suggest that other candidate IGs lack utility, but rather positions IGHV3-64 as a higher-priority carrier candidate within the screened pool. Moreover, the analysis does not imply direct modulation of IRN pharmacodynamics by IGHV3-64; instead, it supports its feasibility as a molecular carrier capable of stable drug association—an essential prerequisite for rational PDC design.

Epitope-pairing analysis revealed that IGHV3-64-derived peptides exhibit spatial compatibility with functionally relevant regions of K-Ras, including residues within the P-loop and switch II domains—structural hotspots implicated in CRC oncogenesis. These in-silico findings are biologically plausible in advanced CRC, where chronic exposure to mutant K-Ras-derived neoantigens may promote B cell clonal selection toward K-Ras-reactive IGHV configurations. Crucially, the present study does not propose that IGHV3-64 binding inhibits K-Ras GTPase activity or suppresses downstream RAS/RAF signaling. Instead, the biological significance of this interaction is best understood as a tumor-immune interface phenomenon, reflecting endogenous humoral responses shaped by oncogenic K-Ras signaling. Epitope pairing was therefore intentionally evaluated to assess spatial compatibility and proximity potential, rather than enzymatic inhibition.

With respect to specificity, epitope pairing was assessed at the level of sequence and structural compatibility rather than competitive binding or mutational discrimination. While the observed pairing suggests preferential interaction, cross-reactivity with structurally related epitopes cannot be excluded. Accordingly, epitope pairing is framed here as candidate neoantigen recognition, not definitive specificity. Future studies should incorporate mutant vs. wild-type K-Ras comparisons and competitive binding assays [e.g., surface plasmon resonance (SPR), co-immunoprecipitation (CoIP), Förster resonance energy transfer (FRET)] to rigorously evaluate selectivity. Within this proximity-guided delivery model, IGHV3-64-derived peptides may act as molecular guides, facilitating localized accumulation of conjugated therapeutic payloads in K-Ras-driven tumor cells. This approach aligns with emerging antibody-guided and peptide-based strategies designed to address the long-standing challenge of targeting “undruggable” oncogenes without requiring direct catalytic inhibition [19].

Additionally, the identification of intrinsic CPP motifs within IGHV3-64 strengthens the biological plausibility of this model by suggesting enhanced intracellular accessibility. Collectively, drug association, epitope proximity, and CPP features define a multifunctional but non-inhibitory role for IGHV3-64 in PDC systems, consistent with engineered antibody-peptide platforms targeting intracellular oncogenic or immune signaling pathways [20].

This study represents the first systematic exploration of stage-specific, stable, secreted circulating IGs shaped by the CRC immune microenvironment and their feasibility as endogenous carriers for drug conjugation, including their association with K-Ras. From a translational perspective, several practical considerations are relevant for IGHV3-64-based pPDC development, including conjugation strategy, stability, and pharmacokinetics. Conceptually, pPDCs may be generated through direct chemical conjugation, linker-based attachment, or emerging linker-free approaches such as fusion tags or site-specific antibody modification. While endogenous IG frameworks may offer intrinsic stability and immunological compatibility, factors such as drug-to-antibody ratio heterogeneity, batch variability, and pharmacokinetic behavior will require systematic optimization and validation. These aspects are acknowledged here as developmental challenges rather than resolved advantages, reinforcing the exploratory nature of the present work.

In comparison with conventional monoclonal ADCs, pPDCs are not proposed as replacements but as complementary modalities. While ADCs benefit from precise epitope targeting and established clinical pipelines, they are limited by single-epitope dependence, antigen loss, immunogenicity, manufacturing cost, and reduced efficacy in antigen-heterogeneous tumors. pPDCs, by contrast, offer multi-epitope coverage and immune-context-adapted targeting, which may be particularly relevant in advanced-stage and K-Ras-mutant CRC, where effective ADC targets and anti-EGFR therapies are limited [21]. Clinically, IRN remains a cornerstone of metastatic CRC therapy, particularly in folinic acid, FU, and IRN (FOLFIRI)-based regimens. However, systemic toxicity, therapeutic resistance, and limited selectivity remain significant challenges. In K-Ras-mutant CRC, where EGFR-targeted antibodies are ineffective, IGHV3-64-based pPDCs may offer a conceptually distinct and complementary strategy, integrating chemotherapeutic delivery with tumor- and immune-context-specific molecular guidance rather than competing with existing regimens [22].

Beyond therapeutic considerations, circulating biomarkers are increasingly explored for prognostic and perioperative risk stratification in CRC. Recent studies highlighting serum butyrylcholinesterase (BuChE) as a predictor of postoperative complications further underscore the clinical relevance of blood-based biomarkers. In this context, stage-specific IG profiles identified in the present study may represent an additional layer of biologically informed biomarkers with potential relevance across diagnostic, prognostic, and therapeutic domains [23].

Collectively, our findings propose a biologically grounded, hypothesis-driven framework for a new class of pPDCs leveraging tumor-associated circulating IGs as polyclonal, multi-epitope carriers. While preliminary and computational in nature, this framework is particularly relevant for ASCRC, where immune remodeling, antigenic heterogeneity, and therapeutic resistance converge. Future studies will focus on experimental validation in vitro and in vivo, including CRC organoid-based pharmacokinetics, tumor selectivity, immune interactions, diagnostic performance (ROC/AUC), and targeted validation using SPR, CoIP, and FRET assays. To clarify the conceptual positioning of pPDCs relative to established ADCs, a comparative overview of key features, advantages, and limitations is provided in Table 7.

This study presents a first-of-its-kind, integrative proteomic and computational framework for identifying circulating, stage-specific IGs in CRC and evaluating their feasibility as endogenous carriers for pPDCs. Selected IGs, particularly IGHV3-64, demonstrated stable association with IRN and epitope-level spatial compatibility with K-Ras, supported by the presence of intrinsic CPP motifs that may facilitate intracellular access. Rather than establishing therapeutic efficacy, these findings identify IGHV3-64 as a candidate IG scaffold with properties suitable for proximity-guided drug delivery in advanced and treatment-resistant CRC. The results support the concept that circulating, immune-selected IGs—shaped by the TME—can be systematically mined and repurposed as immunologically compatible, precision-guided delivery frameworks, pending rigorous experimental validation.

Although primarily exploratory and computational in nature, this work provides a biologically grounded rationale for future in vitro and in vivo studies, including organoid-based models, pharmacokinetic analyses, and functional validation assays. Collectively, the study establishes a conceptual foundation for immune-context-aware precision therapeutic design, offering a new avenue for integrating tumor immunobiology with next-generation drug conjugation strategies in CRC.

A key strength of this study is its integrative, multi-omics approach—combining high-resolution mass spectrometry, computational docking, and epitope-level biophysical analysis. This strategy enabled the identification of stage-specific IG subtypes in CRC and their evaluation as endogenous drug carriers. Unlike conventional ADCs, the use of naturally circulating IGs offers a personalized and immunocompatible platform, potentially minimizing off-target toxicity and improving bioavailability. Notably, the identification of dual-function peptides—with both drug-binding and cell-penetrating capabilities—presents an innovative solution for intracellular delivery, especially for targeting traditionally “undruggable” protein K-Ras.

However, several limitations warrant consideration. First, the findings are based primarily on in-silico and in vitro analyses, necessitating further validation in cellular and animal models to confirm therapeutic relevance. Second, although IG expression patterns were derived from plasma proteomics, it remains unclear whether these signatures are consistent across diverse patient populations or influenced by variables such as immune status, comorbidities, or the gut microbiome. Moreover, the broader functional role of circulating IGs in tumor progression and immune modulation was not investigated, which may impact their utility as therapeutic vectors.

Future work will focus on validating IGs-drug interactions using SPR and assessing cellular uptake. In vivo studies in CRC mouse models will be crucial to evaluate pharmacokinetics, biodistribution, tumor specificity, and therapeutic efficacy. Additional efforts will explore engineering IG-derived peptides with cleavable linkers and payloads such as siRNAs or immunomodulators, paving the way for next-generation pPDCs tailored to heterogeneous and treatment-resistant CRC subtypes.