MAP Crosslinker Chirality Regulates Macrophage Polarization Through MD‐2 Engagement and Enables Tunable Immunomodulation Through Novel Peptide Design

ABSTRACT

Microporous annealed particle (MAP) scaffolds are injectable hydrogel biomaterials that promote tissue regeneration by enabling rapid cell infiltration and presenting reparative cues. Previous work showed that substituting L‐ with D‐chiral residues in matrix metalloproteinase (MMP)‐degradable crosslinkers induces adaptive immune‐mediated skin regeneration and balanced macrophage responses in vivo, but the underlying mechanisms remain unclear. Here, we identify myeloid differentiation factor 2 (MD‐2)‐associated TLR4 signaling as a key mechanistic contributor to macrophage polarization driven by crosslinker chirality in MAP scaffolds. Macrophages cultured in D‐chiral MAP (DMAP) scaffolds exhibited reduced iNOS, CD86, and TNF‐α expression and shifted from an M1‐like state toward M0‐like phenotypes compared to L‐chiral MAP (LMAP) and 2D controls. Competitive inhibition with soluble D‐chiral MMP crosslinker peptide (DMMP) partially restored M1 activation, implicating direct peptide–macrophage interactions. Docking and surface plasmon resonance (SPR) analyses revealed higher‐affinity binding of DMMP to MD‐2 relative to L‐chiral peptides. Accordingly, DMAP altered TLR4/MD‐2 trafficking and attenuated endocytosis‐dependent signaling. Engineering a second D‐peptide crosslinker with enhanced MD‐2 affinity (DMMP2) reduced inflammatory markers and promoted regenerative macrophage polarization. Together, these results establish peptide chirality as a tunable design parameter for modulating TLR4/MD‐2 signaling and engineering immunomodulatory MAP scaffolds.

D‐chiral peptide crosslinkers within microporous annealed particle (MAP) scaffolds preferentially engage the TLR4 co‐receptor MD‐2, altering receptor trafficking and downstream inflammatory signaling while preserving receptor‐proximal innate immune sensing. This mechanistic insight enables affinity‐guided engineering of peptide crosslinkers that tune macrophage phenotype and support regenerative tissue repair.

1. Introduction

Macrophages are a central component of biomaterial–host interactions, integrating chemical and physical cues to regulate inflammation and tissue repair [1, 2, 3, 4]. Their pronounced phenotypic plasticity makes them a key design target for regenerative biomaterials, as macrophages span a continuum of inflammatory and reparative states often simplified to M1 (pro‐inflammatory) and M2 (pro‐regenerative) phenotypes [5].

Microporous annealed particle (MAP) scaffolds, a class of granular hydrogels formed by covalently annealed hydrogel microparticles (HMPs), have emerged as potent immunomodulatory materials [6] capable of supporting macrophage infiltration, promoting organized matrix deposition [7, 8], and enhancing repair across multiple tissues [9].

A defining feature of MAP scaffolds is their tunability through peptide crosslinker chemistry. The individual HMPs that make up MAP scaffolds can be crosslinked through MMP‐degradable peptide sequences to enable cell‐mediated degradation and integration with surrounding tissues [10]. Early work demonstrated that the regenerative potential of MAP scaffolds can be dramatically altered by modifying the chirality of residues within these peptide crosslinkers. Our previous work shows that replacing the canonical L‐chiral MMP‐degradable sequence (LMMP) with a D‐chiral variant containing select D‐amino acids (DMMP) produced striking regenerative outcomes, including de novo hair follicle formation and reduced scarring [11]. Subsequent subcutaneous analyses of immune infiltration across multiple time points revealed that macrophages are the dominant infiltrating immune population within MAP implants. Importantly, crosslinker chirality altered the macrophage response, with DMAP scaffolds promoting distinct macrophage populations co‐expressing both pro‐inflammatory and pro‐regenerative markers, particularly at day 7, compared to their LMAP counterparts. This macrophage profile was strongly associated with the pro‐regenerative outcomes observed in DMAP [12].

Most recently, proteome‐wide temporal profiling of wound healing under LMAP versus DMAP treatments revealed that chirality fundamentally alters both immune and tissue‐level healing trajectories [8]. Across multiple time points, DMAP consistently enriched early inflammatory and immune‐recruitment pathways, followed by a transition toward keratinization, lipid oxidation, and muscle‐associated processes. In contrast, LMAP promoted accelerated transition into proliferative and ECM‐organizing pathways [8]. These findings reinforce that DMAP uniquely reshapes the inflammatory landscape and healing trajectory rather than simply accelerating conventional repair.

Despite the reproducible and powerful regenerative effects of DMAP scaffolds, the mechanistic basis for chirality‐dependent macrophage modulation is unknown. The large regenerative differences that result from these altered crosslinkers, like delayed material degradation, do not sufficiently explain the macrophage‐specific phenotypes or the early immune activation seen in vivo. No work has determined whether D‐chiral peptides directly engage macrophage receptors, whether they alter specific signaling pathways, or how such interactions might redistribute macrophage phenotypes within the heterogeneous populations present in MAP scaffolds.

These unresolved questions underscore a critical gap: we currently lack a mechanistic framework linking peptide‐level design choices to macrophage activation states and downstream tissue repair outcomes. Understanding how crosslinker residue chirality influences macrophage signaling would not only clarify the origins of DMAP's regenerative effects but also open the door to designing crosslinkers that tune macrophage behavior and tissue regeneration in predictable ways.

In this study, we address this gap by investigating how peptide crosslinker chirality regulates macrophage phenotype at a mechanistic level. Through macrophage phenotyping, receptor‐binding studies, and engineered crosslinker variants, we identify MD‐2 associated TLR4 signaling pathway as an important modulatory mechanism underlying chirality‐dependent macrophage responses. Together, these findings establish crosslinker engineering as a powerful strategy for tuning MAP scaffold immunomodulation and regenerative outcomes.

2. Results and Discussion

2.1. Crosslinker Chirality Modulates Macrophage Activation States Away From Inflammatory Phenotypes

We first evaluated MAP scaffold formulations designed to isolate crosslinker chirality as the primary material variable. LMAP and DMAP scaffolds were fabricated using the same MAP chemistry but incorporated MMP‐sensitive crosslinking peptides composed of L‐ or D‐chiral residues, respectively. Microgels exhibited average diameters within the typical MAP scaffold range of approximately 50–100 µm, with no significant differences observed among LMAP and DMAP, formulations (Figure S1). Previously reported characterization of scaffolds fabricated using the same MAP chemistry demonstrated comparable void fractions (∼20%–30%), storage moduli (G′ ≈ 600–800 Pa), and loss moduli (G″ ≈ 20–60 Pa) between formulations [12]. Thus, the primary design variable among scaffold groups was the chirality of the MMP‐sensitive crosslinking peptide.

After establishing the material design, we examined bone marrow‐derived macrophages (BMDMs) cultured in 2D, LMAP, or DMAP scaffolds under M1‐polarizing conditions using flow cytometry and ELISA (Figure 1A and Table S1). BMDMs were cultured in 2D, LMAP, or DMAP conditions for 24 h, polarized with an M1‐activating cytokine cocktail (20 ng/mL LPS and IFN‐γ), and harvested after an additional 24 h for flow cytometry and ELISA.

FIGURE 1.

MAP scaffold crosslinker chirality modulates macrophage phenotypic responses following LPS/IFNγ (M1) polarization. (A) Flow cytometry gating strategy for macrophage identification. Cells were gated by size and granularity with single‐cell (singlet) exclusion using FSC‐A versus SSC‐A, followed by viability gating using Zombie NIR. Macrophages were identified as CD11b⁺ F4/80⁺ cells, and expression of activation markers was quantified using fluorescence intensity histograms. (B) Quantification of macrophage polarization markers across 2D, LMAP, and DMAP conditions. Pro‐inflammatory markers (iNOS, CD86, and TNF measured by ELISA) were significantly elevated in 2D cultures relative to LMAP and DMAP. Antigen‐presenting markers MHCII and CD11c were decreased considerably in MAP‐treated macrophages, with DMAP showing the lowest MHCII expression. (C) sUMAP visualization of eight macrophage clusters reveals distinct phenotypic groupings across biomaterial conditions, ranging from M1‐activated, APC‐like, transitional, and semi‐mature APC‐like states. (D) Heatmap summarizing marker expression among clusters. Cluster annotations were assigned descriptively following FlowSOM clustering based on marker‐expression patterns. M1‐like clusters exhibited high iNOS and CD86 expression, transitional clusters displayed intermediate marker expression, and M0‐like clusters exhibited low expression of all markers. (E) Distribution of macrophage subsets across treatment conditions. Both MAP scaffolds reduced the frequency of M1‐activated macrophages (cluster 1) and increased M0‐like (cluster 2) and CD11c⁻MHCII‐early M1 (cluster 3) populations compared to 2D cultures, with DMAP inducing the largest shifts. Data are presented as mean ± S.E.M. Statistical comparisons were performed using one‐way ANOVA and two‐way ANOVA with Tukey's post hoc test; p‐values are indicated above bars. n = 3 independent experiments.

MAP scaffolds broadly attenuated inflammatory activation relative to 2D controls, consistent with previous observations [13], but DMAP elicited the most pronounced suppression. Expression of M1 markers, including iNOS, CD86, and secretion of a pro‐inflammatory cytokine, TNF, was significantly lower in DMAP‐treated macrophages compared to both LMAP and 2D cultures, indicating that D‐chiral crosslinkers more effectively suppress pro‐inflammatory signaling than their L‐chiral counterparts. Antigen‐presentation markers such as MHCII and CD11c were similarly reduced in MAP cultures, with DMAP yielding the lowest overall expression (Figure 1B).

To further evaluate how crosslinker chirality reshapes macrophage population composition, we performed FlowSOM clustering and visualized the resulting populations using supervised UMAP (sUMAP). This analysis resolved into eight distinct macrophage subsets (Figure 1C,D). Clusters with high iNOS and CD86 expression were classified as M1‐like populations, clusters exhibiting intermediate activation‐marker expression were designated transitional M1‐like populations, and clusters displaying comparatively low expression of all markers were designated M0‐like populations. Macrophages cultured on 2D substrates, and to a lesser extent in LMAP scaffolds, were dominated by the M1‐like cluster. In contrast, DMAP cultures exhibited a significant redistribution away from these inflammatory states toward the M0‐like subsets. Quantification of cluster frequencies confirmed that DMAP significantly reduced the abundance of the dominant M1 cluster, decreasing from more than 60% in 2D cultures to less than 50% in DMAP, while increasing the representation of M0‐like populations (exceeding 20%) as well as early antigen‐presenting M1‐like macrophages (Figure 1E).

To determine whether the presence of chiral peptides alone is sufficient to induce this phenotypic shift, we compared the effects of soluble LMMP and DMMP peptides on 2D macrophage cultures under M1‐polarizing conditions with those observed in their scaffold‐based counterparts (LMAP and DMAP). As previously reported [11], the addition of soluble DMMP failed to attenuate M1 marker expression, yielding activation profiles similar to untreated macrophages and remaining significantly elevated, particularly for iNOS and antigen‐presentation markers, compared to DMAP cultures (Figure S2A). These findings indicate that the presence of D‐chiral peptides in solution is insufficient to recapitulate the immunomodulatory effects of DMAP scaffolds and that peptide presentation within the MAP scaffold architecture is required to dampen pro‐inflammatory activation.

Given that the macrophage phenotype is known to be influenced by material‐derived cues such as spatial confinement [14], surface modification [15], and cell shape [16], we next analyzed macrophage morphology in both LMAP and DMAP scaffolds under M0, M1, or M2 polarizing conditions using IMARIS. Consistent with established literature, M1 macrophages typically exhibit a more flattened and enlarged shape, whereas M2 macrophages tend to adopt a more elongated shape [13]. Confocal imaging and three‐dimensional cell renderings revealed no significant differences in macrophage morphology between LMAP and DMAP scaffolds across polarization states (Figure S2B,C), demonstrating that both scaffolds provide comparable physical and geometric cues. These data suggest that the reduced pro‐inflammatory response observed in DMAP is not driven by differences in macrophage morphology or surface interactions but instead arises from a distinct chirality‐dependent signaling mechanism.

In summary, DMAP scaffolds suppress classical M1 activation and bias macrophages toward minimally activated M0‐like states. This chirality‐dependent shift in macrophage phenotype suggests that direct interactions between D‐peptides and macrophage receptors may contribute to the observed immunomodulatory effects, motivating future mechanistic studies.

2.2. Competitive Inhibition by Soluble DMMP Reveals Direct Peptide–Receptor Interactions Impacting Macrophage Polarization

Macrophage responses to DMAP scaffolds could arise either from direct recognition of D‐peptide crosslinker motifs at the cell surface or from indirect effects on the local microenvironment, such as altered protein adsorption or cytokine sequestration [17]. Discriminating between these possibilities is essential to determine whether DMAP acts through a receptor‐mediated mechanism that can be rationally tuned through peptide design. To test whether D‐chiral crosslinker peptides directly engage macrophage receptors to modulate activation, we performed competitive inhibition studies by adding soluble DMMP peptide during DMAP scaffold culture. If DMAP‐mediated immunomodulation arises from receptor‐level peptide binding, then soluble DMMP would compete with scaffold‐bound D‐chiral crosslinkers for receptor engagement and partially restore inflammatory signaling (Figure 2A).

FIGURE 2.

Free DMMP peptide competitively disrupts DMAP‐macrophage interactions and partially restores pro‐inflammatory macrophage activation. (A) Schematic illustrating macrophage interactions with DMAP scaffolds under normal conditions (left) and during competitive binding with free DMMP peptide (right). In the absence of soluble DMMP, macrophages directly engage DMMP motifs presented on DMAP microgels. Addition of free DMMP competitively occupies ligand‐binding interactions in solution, reducing macrophage‐DMAP engagement. (B) TNF secretion measured by ELISA from macrophages cultured in 2D, DMAP, or DMAP in the presence of increasing concentrations of free DMMP. DMAP significantly suppressed TNF relative to 2D controls, and addition of free DMMP partially restored TNF secretion toward 2D levels. (C) Quantification of macrophage polarization markers through flow cytometry across LMAP, DMAP, and DMAP + DMMP conditions. Pro‐inflammatory markers (iNOS, CD86) were significantly increased in LMAP compared to DMAP and DMAP+DMMP (1:4 DMAP: Soluble DMMP). CD86 expression was also significantly higher in DMAP+DMMP than in DMAP. Statistical differences between treatments are indicated above each plot. (D) sUMAP projection of seven macrophage clusters across LMAP, DMAP, and DMAP+DMMP conditions. (E) Heatmap summarizing relative expression of Arg1, CD206, CD11c, CD86, MHCII, and iNOS across seven macrophage clusters, revealing a spectrum of phenotypes ranging from M0‐like and transitional APC‐like states to antigen‐presenting and fully mature M1 macrophages characterized by high CD86 and iNOS expression. (F) Distribution of macrophage subsets across treatments. DMAP and DMAP+DMMP decreased the proportion of one M1‐fully mature macrophage cluster (cluster 7) relative to LMAP. However, free DMMP increased the proportion of the transitional APC‐like M1 relative to both DMAP and LMAP (cluster 1), reflecting an overall slight restoration of M1 polarization with the addition of free DMMP. Data are presented as mean ± S.E.M. Statistical comparisons were performed using one‐way ANOVA and two‐way ANOVA with Tukey's post‐hoc test; p values are indicated above bars. n = 3 biological replicates (BMDMs independently isolated from three separate mice).

Consistent with this hypothesis, the addition of soluble DMMP induced a dose‐dependent increase in macrophage activation. TNF secretion, which was strongly suppressed in DMAP cultures relative to 2D controls, progressively increased upon addition of free DMMP, trending back toward levels observed in inflammatory conditions (Figure 2B). Flow cytometry assessment further supported this interpretation, as expression of iNOS and CD86 remained lowest in DMAP‐treated macrophages but increased significantly when soluble DMMP was added (1:4 DMAP: soluble DMMP). Notably, iNOS and CD86 expression exhibited an intermediate inflammatory phenotype between DMAP alone and LMAP, indicating partial restoration of M1‐like activation (Figure 2C). Together, these results support a competitive binding model in which soluble DMMP interferes with macrophage engagement of scaffold‐bound D‐peptide motifs.

To determine whether this effect was specific to D‐chiral peptides, we performed a parallel competition experiment using soluble LMMP at the same concentration. In contrast to DMMP, soluble LMMP produced only modest changes in pro‐inflammatory marker expression, with a significantly weaker increase in iNOS and CD86 compared to DMMP supplementation (Figure S3). These results indicate that D‐chiral peptides more effectively compete with DMAP for macrophage receptor engagement than their L‐chiral counterparts, consistent with a chirality‐dependent interaction.

High‐dimensional profiling further supported a competitive mechanism (Figure 2D.) sUMAP analysis identified seven macrophage populations based on their relative expression of inflammatory, antigen‐presentation, and reparative markers. Among these, cluster 1 was characterized by intermediate CD11c and CD86 expression together with elevated MHCII and iNOS, consistent with a transitional APC‐like M1 phenotype. In contrast, cluster 2 exhibited generally low expression of inflammatory, antigen‐presentation, and reparative markers and was therefore classified as an M0‐like/weakly activated macrophage population (Figure 2E). Relative to LMAP cultures, DMAP‐treated macrophages exhibited a reduction in cluster 1 and a corresponding enrichment of cluster 2, consistent with attenuation of inflammatory activation. Upon addition of soluble DMMP, the abundance of cluster 1 increased and approached levels observed in LMAP cultures, while the M0‐like/weakly activated cluster 2 decreased. These shifts indicate partial restoration of macrophage activation following competitive inhibition of scaffold‐associated DMMP, further supporting a role for direct peptide‐mediated signaling in the observed macrophage phenotype (Figure 2F).

Collectively, these findings support receptor‐level interactions between D‐peptide crosslinkers presented within DMAP scaffolds and macrophage, as evidenced by the competitive inhibition of DMAP‐mediated effects by soluble DMMP. The partial restoration of M1 activation upon soluble DMMP addition suggests that receptor‐level binding events contribute to DMAP‐induced macrophage modulation. The greater competitive efficacy of DMMP relative to LMMP further suggests the involvement of a macrophage receptor that preferentially recognizes D‐chiral peptides, motivating systematic identification of candidate receptors in subsequent analyses.

2.3. In Silico Docking and SPR Binding Results Demonstrate Preferential MD‐2 Binding by DMMP Over LMMP

Given that soluble DMMP partially restored BMDM pro‐inflammatory activation, we next sought to identify the receptor(s) through which DMMP exerts its immunomodulatory effects. Because the effects of DMAP were observed in macrophages driven toward M1 polarization using LPS and IFN‐γ, we hypothesized that DMMP may directly interact with cell‐surface proteins associated with canonical M1‐polarizing pathways (Figure 3A).

FIGURE 3.

LMMP and DMMP differentially interact with MD‐2, which is involved in LPS‐mediated macrophage polarization. (A) Schematic summarizing macrophage surface receptors involved in M2‐polarizing (IL‐4) and M1‐polarizing (LPS/IFN‐γ) signaling. Receptors implicated in potential peptide‐protein interactions, including TLR4 and CD14, are shown for reference. (B) HDOCK analysis interaction scores for LMMP and DMMP. Both peptides demonstrated predicted interactions with TLR4/MD‐2 and CD14, but DMMP demonstrated a consistently higher confidence and a larger binding difference relative to LMMP for MD‐2, suggesting preferential MD‐2 engagement. (C) Protein–peptide docking models illustrating the predicted binding of LMMP (yellow) and DMMP (purple) with chiral residues (orange) within the MD‐2 (green) hydrophobic pocket. Representative zoomed views highlight MD‐2 amino acid residues located within 4 Å of the bound peptides, indicating direct molecular interactions. (D) Proposed mechanism illustrating how differential peptide‐MD‐2 interactions could influence LPS sensing. In the LPS‐triggered M1 polarization pathway, the LPS‐LBP complex delivers LPS to CD14 and subsequently to MD‐2, promoting TLR4/MD‐2 complex endocytosis and downstream signaling. Peptide binding (right) may interfere with LPS sensing for MD‐2 engagement, with DMMP predicted to exhibit stronger interference than LMMP based on HDOCK results. (E) SPR sensorgrams for LMMP, DMMP, and control LPS binding to immobilized MD‐2. Sensorgrams shown represent the complete set of analyte concentrations used for affinity analysis. DMMP produced a larger binding response across tested concentrations relative to LMMP, consistent with higher MD‐2 affinity. (F) Linear regression analysis of SPR equilibrium binding responses for LMMP and DMMP interacting with immobilized MD‐2. Response units (RU) at equilibrium were plotted as a function of analyte concentration, and linear fits were used to estimate apparent dissociation constants (K _D), χ ² values, and R _max. DMMP exhibited a steeper slope and lower apparent K_D than LMMP, indicating stronger MD‐2 binding affinity. (G) Summary of SPR‐derived binding affinities for LPS, LMMP, and DMMP. LPS exhibited submicromolar affinity, while DMMP displayed markedly higher affinity for MD‐2 than LMMP, supporting preferential MD‐2 engagement by DMMP.

To evaluate this possibility, we performed structure‐based in silico docking analyses using HDOCK [18], comparing the predicted interactions of LMMP and DMMP crosslinker peptides with components of both M1‐ and M2‐associated signaling pathways, including TLR4, MD‐2, CD14, the TLR4/MD‐2 complex, IL‐4R, and IL‐10R [19]. Because DMMP is not a naturally occurring peptide, its three‐dimensional structure (.PDB file) (Data S1) was generated using PEPstrMOD [20] prior to docking. Binding likelihood was assessed using HDOCK confidence scores. According to HDOCK guidelines, scores > 0.7 indicate high‐confidence binding, scores 0.5–0.7 indicate possible binding, and scores < 0.5 indicate low likelihood of interaction [18].

Both LMMP and DMMP were predicted to interact with many of the evaluated cell‐surface proteins. However, DMMP consistently demonstrated higher‐confidence binding to the MD‐2 co‐receptor and to CD14 compared to LMMP, with the largest confidence score difference observed for MD‐2 (Figure 3B and Data S2). Docking models further revealed distinct binding modes, with LMMP interacting with 24 MD‐2 residues, of which 9 were located within or near the lipid A binding pocket [21], whereas DMMP interacted with 34 MD‐2 residues, including 15 residues within or proximal to the lipid A binding region [21] (Figure 3C). These results suggest that DMMP occupies a larger and more functionally relevant surface within MD‐2, potentially interfering with ligand recognition or receptor activation.

MD‐2 (myeloid differentiation factor 2) is an essential TLR4 co‐receptor that binds lipid A and other ligands, enabling TLR4 dimerization, receptor endocytosis, and initiation of MyD88‐ and TRIF‐dependent signaling cascades that drive NF‐κB activation and pro‐inflammatory cytokine production [21, 22, 23, 24, 25]. The computational preference of DMMP for MD‐2 over other receptor candidates, combined with MD‐2's central role in TLR4 signaling, positioned the MD‐2–TLR4 axis as the most likely mediator of chirality‐dependent macrophage responses. Accordingly, we hypothesized that stronger engagement of MD‐2 by DMMP presented within DMAP scaffolds disrupts TLR4 signaling and dampens pro‐inflammatory activation (Figure 3D).

We further validated these in silico predictions using surface plasmon resonance (SPR). MD‐2 was immobilized on a CM5 sensor chip, and increasing concentrations of LPS (positive control), LMMP, and DMMP were injected. DMMP produced substantially larger binding responses than LMMP across all tested concentrations, consistent with a higher binding affinity for MD‐2 (Figure 3E). As expected, LPS exhibited the strongest interaction with MD‐2, achieving sub‐micromolar affinity (K _D = 9.10 × 10^{− 7} M), in agreement with previous reports [26, 27]. Steady‐state analysis revealed that DMMP bound MD‐2 with measurable affinity (K _D = 3.60 × 10^{− 2} M), whereas LMMP displayed substantially weaker binding (K _D = 1.72 × 10² M) (Figure 3E–G). Because both LMMP and DMMP exhibited rapid association and dissociation kinetics, traditional kinetic fitting was poorly constrained. Therefore, K _D values were determined using steady‐state response‐versus‐concentration analyses, an approach recommended for low‐affinity, fast‐kinetic interactions in SPR measurements [28, 29].

Although DMMP exhibited measurable binding to MD‐2, its affinity was substantially weaker than that of LPS, supporting a model in which soluble DMMP does not directly compete with LPS for MD‐2 binding. However, the biological activity of scaffold‐presented DMMP may reflect presentation‐dependent effects. In the SPR experiments, soluble peptide binding was measured against immobilized MD‐2, which does not reproduce the multivalent scaffold–cell interface. MAP scaffolds present many copies of DMMP across a cell‐contacting surface, which could increase local peptide availability and enable repeated or multivalent engagement of membrane‐associated TLR4/MD‐2 complexes. Multivalent ligand presentation is known to enhance functional avidity through ligand density, receptor recruitment, and rebinding [30, 31]. Therefore, despite its relatively weak monovalent affinity compared to LPS‐MD‐2, it is plausible that DMMP's preferential binding to MD‐2 compared with LMMP may contribute to modulation of TLR4/MD‐2 signaling and downstream macrophage behavior when presented within the scaffold microenvironment.

Together, the HDOCK and SPR results demonstrate that DMMP peptides directly engage the MD‐2 co‐receptor with substantially higher affinity than LMMP, providing a plausible molecular basis for the chirality‐dependent immunomodulatory effects of DMAP scaffolds. Preferential MD‐2 binding by DMMP supports MD‐2 as a key mechanistic contributor through which crosslinker chirality modulates macrophage activation. Given the central role of MD‐2 in regulating TLR4 trafficking and signaling, these findings further suggest that DMMP–MD‐2 interactions may alter TLR4/MD‐2 localization and endocytic dynamics, motivating direct examination of receptor distribution and endocytosis in subsequent experiments.

2.4. DMAP Modulates MD‐2‐Associated TLR4 Trafficking and Downstream Inflammatory Signaling

M1 activation of macrophages by LPS requires not only binding of LPS to MD‐2 at the cell surface but also subsequent endocytosis of the TLR4/MD‐2 complex, which enables TRIF‐dependent inflammatory signaling [32, 33, 34]. To evaluate whether MD‐2 signaling contributes to the reduced inflammatory phenotype observed in DMAP‐treated macrophages, BMDMs cultured under 2D, LMAP, or DMAP conditions were treated with MD2‐IN‐1, a selective small‐molecule inhibitor of MD‐2, prior to M1 activation. Pharmacologic inhibition of MD‐2 markedly reduced iNOS expression in both 2D and LMAP cultures, driving macrophages toward a phenotype resembling that observed in DMAP‐treated cells (Figure 4A). In contrast, iNOS expression in DMAP‐treated macrophages was already substantially reduced prior to MD2‐IN‐1 treatment, and pharmacologic inhibition produced a further decrease. These findings support a functional role for MD‐2‐associated signaling in the pro‐inflammatory phenotype and suggest that DMAP partially recapitulates the effects of pharmacologic MD‐2 inhibition.

FIGURE 4.

DMAP modulates MD‐2‐dependent TLR4 trafficking and downstream inflammatory signaling. (A) Effects of pharmacologic MD‐2 inhibition on macrophage activation. BMDMs cultured under 2D, LMAP, or DMAP conditions were stimulated under M1‐polarizing conditions in the presence or absence of the MD‐2 inhibitor MD2‐IN‐1. MD2‐IN‐1 markedly reduced iNOS expression in 2D and LMAP cultures, while producing comparatively modest additional effects in DMAP‐treated macrophages, supporting a role for MD‐2‐dependent signaling in the DMAP phenotype. (B) Quantification of intracellular, extracellular, and intra/extracellular ratios of MTS510‐detectable TLR4/MD‐2 complexes in macrophages cultured in 2D, LMAP, or DMAP conditions. DMAP significantly reduced intracellular MTS510‐detectable TLR4/MD‐2 abundance while increasing extracellular signal. The resulting intra/extracellular ratio was markedly lower in DMAP‐treated macrophages, indicating altered distribution of MTS510‐detectable TLR4/MD‐2 complexes consistent with altered receptor trafficking and/or signaling state. C) Effects of endocytosis inhibition on macrophage activation and TLR4/MD‐2 distribution. In LMAP‐treated macrophages, dynasore reduced pro‐inflammatory marker expression (iNOS and CD86) to levels comparable to those observed in DMAP‐treated cells. Notably, DMAP‐treated macrophages exhibited minimal additional changes upon dynasore treatment, whereas LMAP + dynasore closely phenocopied the DMAP condition. Similarly, dynasore altered the distribution of MTS510‐detectable TLR4/MD‐2 complexes in LMAP‐treated macrophages toward a profile resembling that observed in DMAP‐treated cells, supporting a role for altered TLR4/MD‐2 trafficking in the observed immunomodulatory effects. Antigen‐presenting markers (MHCII) were less sensitive to dynasore treatment across conditions. (D) IRF‐Lucia reporter activity was measured in THP1‐dual cells cultured under 2D conditions or within LMAP and DMAP scaffolds and stimulated with LPS. Reporter activity was quantified at 12 and 24 h post‐stimulation. Both LMAP and DMAP reduced IRF‐associated reporter activity relative to LPS‐stimulated 2D controls, with DMAP producing the greatest suppression. (E) NF‐κB‐SEAP reporter activity was measured from the same THP1‐dual cultures at 12 and 24 h post‐stimulation. DMAP significantly reduced NF‐κB reporter activity relative to both 2D and LMAP conditions. Together, these findings demonstrate that DMAP attenuates NF‐κB‐ and IRF‐associated signaling in a human macrophage‐like reporter system, consistent with altered TLR4/MD‐2 trafficking and reduced inflammatory activation. Data are presented as mean ± S.E.M. Individual points represent biological replicates. Statistical analyses were performed using one‐way or two‐way ANOVA with Tukey's multiple‐comparison test, as appropriate; p‐values are indicated above each graph. For BMDM experiments, biological replicates represent macrophages independently isolated from three separate mice (n = 3). For THP1‐Dual reporter assays, biological replicates represent independent THP1‐Dual cultures (n = 3).

We next investigated whether DMAP altered the trafficking dynamics of the TLR4/MD‐2 complex. Intracellular and extracellular MTS510‐detectable TLR4/MD‐2 complexes were quantified by flow cytometry. Because MTS510 recognizes a conformation‐sensitive epitope within the TLR4/MD‐2 complex, these measurements are interpreted as changes in detectable TLR4/MD‐2 distribution rather than direct measurements of total receptor abundance. Under M1‐polarizing conditions, macrophages cultured on 2D substrates exhibited a predominance of intracellular MTS510‐detectable TLR4/MD‐2 signal. LMAP cultures showed a modest reduction in this intracellular pool; however, DMAP induced a markedly distinct distribution. In DMAP‐treated macrophages, intracellular TLR4/MD‐2 levels were significantly decreased, while extracellular TLR4/MD‐2 abundance was increased, resulting in a substantially lower intracellular‐to‐extracellular ratio (Figure 4B). This shift suggests that engagement of MD‐2 by DMMP alters TLR4/MD‐2 trafficking and reduces receptor internalization.

To determine whether impaired endocytosis mechanistically contributes to DMAP's immunomodulatory effects, we pharmacologically disrupted endocytosis using dynasore, an inhibitor of dynamin, a GTPase required for vesicle scission during endocytosis [35]. In 2D M1‐polarized macrophages, dynasore treatment significantly reduced iNOS and CD86 expression, validating the requirement of endocytosis for sustaining full pro‐inflammatory activation. Dynasore also decreased intracellular MTS510‐detectable TLR4/MD‐2 signal and increased extracellular signal in these controls (Figure S4), consistent with altered TLR4/MD‐2 trafficking dynamics following inhibition of endocytosis.

Dynasore was then applied to macrophages cultured in LMAP and DMAP scaffolds. In LMAP cultures, dynasore significantly reduced iNOS and CD86 expression relative to untreated LMAP controls, phenocopying the anti‐inflammatory effects observed in DMAP. In contrast, dynasore treatment produced minimal additional changes in DMAP‐treated macrophages, where pro‐inflammatory marker expression remained similarly suppressed regardless of endocytosis inhibition (Figure 4C). These findings suggest that DMAP partially recapitulates signaling consequences associated with endocytosis inhibition, rendering additional pharmacological blockade less impactful.

To determine whether the alterations in TLR4/MD‐2 trafficking translated to downstream signaling changes, we next utilized THP1‐Dual reporter cells, a human macrophage‐like cell line that enables simultaneous monitoring of NF‐κB activation through secretion of SEAP and IRF activation through secretion of Lucia luciferase. Because plasma membrane‐localized TLR4 signaling is primarily associated with MyD88‐dependent NF‐κB activation, whereas endosomal TLR4 signaling engages the TRIF–IRF axis [32, 34], this reporter system provides a functional readout of both signaling branches.

Consistent with the trafficking data, LPS‐stimulated THP1‐Dual cells cultured in DMAP exhibited significantly reduced NF‐κB and IRF reporter activity compared with both 2D and LMAP controls (Figure 4D,E). In contrast, LMAP primarily attenuated IRF‐Lucia activity while maintaining NF‐κB‐SEAP activity comparable to LPS‐stimulated 2D cultures. Because IRF activation depends on endosomal TLR4 trafficking and TRIF engagement, the preferential reduction in IRF signaling observed in both MAP scaffold conditions, and particularly in DMAP, is consistent with altered TLR4/MD‐2 trafficking. The broader suppression of both NF‐κB and IRF signaling in DMAP further suggests that D‐chiral crosslinkers attenuate inflammatory signaling downstream of TLR4 activation.

Together, these findings support a functional role for MD‐2‐dependent signaling in the immunomodulatory effects of DMAP. Pharmacologic inhibition of MD‐2 recapitulated key aspects of the DMAP phenotype, while DMAP itself altered the distribution of MTS510‐detectable TLR4/MD‐2 complexes and reduced the intracellular‐to‐extracellular receptor ratio. Similarly, pharmacologic inhibition of endocytosis phenocopied the anti‐inflammatory effects observed in DMAP‐treated macrophages, indicating that altered receptor trafficking contributes to the observed phenotype. Finally, reduced NF‐κB and IRF reporter activity in THP1‐Dual cells provided functional evidence of diminished downstream inflammatory signaling in a human macrophage‐like system. Collectively, these results support a model in which D‐chiral crosslinkers modulate macrophage inflammatory responses through MD‐2‐associated alterations in TLR4 trafficking and signaling, resulting in attenuation of both pro‐inflammatory activation and endocytosis‐dependent signaling pathways.

2.5. Temporal Wound Proteomics Reveals Altered TLR4‐Associated Signaling Programs in DMAP‐Treated Wounds

To investigate whether the signaling changes observed in vitro were reflected within the wound environment, we performed a targeted analysis of a previously published temporal proteomics dataset comparing LMAP‐ and DMAP‐treated wounds at days 4, 14, and 21 post‐wounding [8]. These time points span the major phases of wound healing, including inflammation, proliferation, and remodeling. From the full proteomic dataset, we identified proteins associated with receptor‐proximal TLR/PRR sensing, NF‐κB/MAPK signaling, ROS and metabolic stress responses, and the endocytosis‐dependent TRIF/IRF/type I interferon pathway.

Temporal heatmap analysis revealed that the greatest signaling differences between DMAP and LMAP occurred at day 4, corresponding to the inflammatory phase of wound healing (Figure 5A). At this time point, multiple proteins associated with NF‐κB/MAPK signaling and TRIF/IRF/type I interferon responses were reduced in DMAP‐treated wounds relative to LMAP‐treated wounds. In contrast, proteins involved in receptor‐proximal TLR sensing and adaptor recruitment did not display a consistent pattern of suppression, suggesting that DMAP does not broadly inhibit pathogen‐sensing machinery but instead modulates signaling downstream of receptor engagement. Since these coordinated differences were evident on day 4 but were largely absent at later time points, these data suggest that the effects of DMAP on TLR4‐associated signaling are most pronounced during the early inflammatory stage of wound healing and become less distinct as wounds transition into proliferative and remodeling phases.

FIGURE 5.

Temporal wound proteomics supports altered TLR4‐associated signaling trajectories in DMAP‐treated wounds. A) Heatmap showing targeted reanalysis of a previously published temporal wound proteomics dataset, focusing on proteins associated with NFκB/MAPK signaling, proximal TLR/PRR signaling, ROS/NADPH oxidase activity, and TRIF–STING–IRF/type I interferon signaling. Protein abundance is displayed as row‐normalized Z‐scores across DMAP‐ and LMAP‐treated wounds at days 4, 14, and 21 post‐wounding. (B) Volcano plot comparing DMAP‐ and LMAP‐treated wounds at day 4. Highlighted proteins represent signaling molecules associated with receptor trafficking, inflammatory signaling, innate immune sensing, and interferon‐associated pathways. Positive fold‐change values indicate proteins enriched in DMAP‐treated wounds, whereas negative fold‐change values indicate proteins enriched in LMAP‐treated wounds. C) Quantification of receptor‐proximal TLR4/MD‐2 sensing components, including LBP, CD14, IRAK4, and MYD88. DMAP‐treated wounds maintained expression of ligand‐sensing and proximal signaling proteins, indicating preservation of upstream innate immune recognition at day 4. (D) Quantification of downstream signaling and trafficking regulators, including SRC, SQSTM1, MAP2K1, and PTPN11. DMAP‐treated wounds exhibited reduced abundance of multiple proteins associated with signal propagation, receptor trafficking, and inflammatory pathway regulation. (E) Quantification of endocytosis‐dependent TRIF/IRF signaling proteins, including IFNB1, DDX58, and ISG15. Reduced IFNB1 abundance in DMAP‐treated wounds is consistent with altered TLR4/MD‐2 trafficking and diminished endocytosis‐dependent signaling at day 4. Together, these analyses support a model in which DMAP does not globally suppress innate immune activation but instead rewires TLR4‐associated signaling pathways during wound healing. Data are presented as mean ± S.E.M. Statistical comparisons between LMAP‐ and DMAP‐treated wounds were performed using two‐tailed Welch's t‐tests; p values are indicated above each graph. n = 3–4 wounds per treatment.

To identify the proteins contributing most strongly to the day 4 differences between treatment groups, we generated a volcano plot using the targeted TLR4‐associated protein set at day 4 (Figure 5B). Relative to LMAP, DMAP‐treated wounds exhibited reduced abundance of proteins involved in kinase signaling and signal propagation, including Src, Btk, Sqstm1, Nfkb1, and Stat5a, as well as the type I interferon mediator Ifnb1 [36, 37]. At the same time, DMAP‐treated wounds maintained or increased expression of proteins associated with cellular sensing and stress adaptation, including Ddx58, Gpx4, Cybb, Hck, and Saa4. This pattern suggests that DMAP does not silence innate immune activity but instead redirects signaling away from classical inflammatory activation while preserving pathways involved in cellular sensing and adaptive stress responses.

To relate these observations to specific signaling modules, we next examined representative proteins from distinct stages of the TLR4 pathway at day 4. Receptor‐proximal sensing proteins, including CD14, IRAK4, and MYD88 [38, 39], were largely preserved in DMAP‐treated wounds, with only LBP exhibiting a significant increase (Figure 5C). In contrast, proteins associated with downstream signal propagation and receptor trafficking, including SRC, SQSTM1, MAP2K1, and PTPN11 [36, 40], were significantly reduced (Figure 5D). Most notably, DMAP‐treated wounds exhibited suppression of the endocytosis‐dependent TRIF/IRF signaling axis, characterized by reduced IFNB1 expression together with altered DDX58, ISG15, and STAT2 abundance (Figure 5E).

Collectively, these findings indicate that DMAP does not broadly suppress TLR4 signaling. Rather, DMAP preserves receptor‐proximal sensing while attenuating downstream signaling and endocytosis‐dependent TRIF/IRF activation during the early inflammatory phase of wound healing. This pattern closely mirrors the in vitro observations of altered TLR4/MD‐2 trafficking, reduced NF‐κB and IRF reporter activity, and diminished inflammatory activation, providing pathway‐level evidence that DMAP rewires inflammatory signaling during the initial stages of tissue repair rather than inducing sustained suppression of innate immune responses throughout healing.

2.6. Engineering D‐peptide Crosslinkers With Enhanced MD‐2 Affinity Amplifies Macrophage Immunomodulation

The identification of MD‐2 as a preferential binding target for DMMP suggested that tuning peptide affinity for MD‐2 could serve as a rational design parameter for modulating macrophage responses within MAP scaffolds. To test this hypothesis, we engineered an additional D‐peptide crosslinker, DMMP2, by systematically varying the chirality of residues within the MMP‐degradable recognition motif and screening candidate sequences for predicted MD‐2 interaction strength using HDOCK [18] (Figure 6A,B and Data S4). Among the candidate peptides, DMMP2 (Ac‐GCRDGPQGIW_DGQ_DDRCG‐NH₂) was selected for experimental validation because it combined one of the highest predicted MD‐2 interaction scores while requiring only two D‐amino acid substitutions relative to the LMMP sequence. Unlike DMMP, which contains three D‐residues distributed throughout the recognition sequence, DMMP2 represents an intermediate design that enabled us to evaluate whether enhanced MD‐2 engagement could be achieved through minimal chirality modifications rather than extensive sequence‐wide substitution.

FIGURE 6.

Docking‐guided optimization identifies DMMP2 as a higher‐affinity MD‐2‐binding crosslinker with enhanced immunomodulatory activity. (A) Schematic of the MMP‐degradable crosslinker sequence highlighting the peptide recognition motif and amino acid positions selected for chirality variation. Residues within the recognition sequence were systematically converted from L‐ to D‐chirality to generate a library of candidate peptides. (B) In silico screening of candidate peptide variants. Peptides were ranked according to predicted MD‐2 interaction scores generated using HDOCK. DMMP2 was selected for further study because it exhibited one of the highest predicted MD‐2 interaction scores while requiring only two D‐amino acid substitutions relative to the parent LMMP sequence. (C) Representative docking model of DMMP2 (magenta) bound within the MD‐2 binding pocket (green). Enlarged view highlights predicted interactions between DMMP2 and residues within the MD‐2 pocket. (D) Surface plasmon resonance (SPR) analysis of peptide binding to immobilized MD‐2. Representative sensorgrams and affinity analysis demonstrated that DMMP2 exhibited substantially greater binding affinity for MD‐2 than DMMP and LMMP, consistent with the docking predictions. (E) Functional validation of MD‐2 involvement in the DMAP2 phenotype. BMDMs cultured in DMAP2 scaffolds were treated with the MD‐2 inhibitor MD2‐IN‐1 prior to M1 activation. Pharmacologic inhibition produced only a modest additional reduction in iNOS expression, indicating that DMAP2 already promotes substantial attenuation of MD‐2‐dependent inflammatory signaling. (F) Flow cytometric assessment of macrophage polarization markers following culture in LMAP, DMAP, or DMAP2 scaffolds. DMAP2 further reduced expression of pro‐inflammatory markers (CD86 and iNOS) while increasing expression of the pro‐regenerative marker Arg1 relative to LMAP and DMAP. CD206 expression remained comparable among groups. These findings demonstrate that rational modulation of peptide chirality can enhance MD‐2 engagement and further tune macrophage phenotype toward a pro‐regenerative state. Data are presented as mean ± S.E.M. Individual points represent biological replicates. Statistical analyses were performed using one‐way ANOVA with Tukey's multiple‐comparison test; p values are indicated above each graph. n = 3 biological replicates for Panel D and n = 3–4 biological replicates for Panel F, with BMDMs independently isolated from three separate mice.

DMMP2 exhibited a higher predicted MD‐2 confidence score than DMMP, and docking models revealed more extensive engagement with the MD‐2 hydrophobic pocket [25, 41] (Figure 6A–C and Data S3 and S4). Specifically, DMMP2 was predicted to interact with 23 MD‐2 residues, including 12 residues located within or proximal to the lipid A binding pocket, compared to more peripheral contacts observed for DMMP. These docking predictions were subsequently validated experimentally by SPR, which demonstrated that DMMP2 binds MD‐2 with higher affinity than both DMMP and LMMP (Figure 6D). Steady‐state analysis yielded a K_D of 2.10 × 10^{− 3} M for DMMP2, representing an approximately 17‐fold increase in affinity relative to DMMP and several orders of magnitude stronger binding than LMMP.

To further determine whether the enhanced activity of DMMP2 remained associated with MD‐2‐dependent signaling, macrophages cultured in DMAP2 scaffolds were treated with the MD‐2 inhibitor MD2‐IN‐1 prior to M1 activation. Pharmacologic inhibition produced only a modest additional reduction in iNOS expression relative to untreated DMAP2 cultures (Figure 6E), suggesting that DMAP2 already drives substantial attenuation of MD‐2‐associated inflammatory signaling.

To determine whether this enhanced MD‐2 affinity translated into stronger immunomodulatory effects, we synthesized MAP scaffolds using DMMP2 as the crosslinker peptide to generate DMAP2, and further assessed macrophage activation under M1‐polarizing conditions, as in prior studies. Flow cytometry analysis revealed a clear affinity–phenotype relationship where the expression of the pro‐inflammatory markers iNOS and CD86 was lowest in DMAP2‐treated macrophages, intermediate in DMAP, and highest in LMAP cultures (Figure 6F).

Beyond suppressing pro‐inflammatory signaling, DMAP2 further promoted pro‐reparative macrophage features. Expression of Arg1, which trended upward in DMAP relative to LMAP, was significantly increased in DMAP2‐treated macrophages, indicating an enhanced bias toward a reparative phenotype (Figure 6F). CD206 expression followed a similar upward trend, further supporting a shift toward regenerative macrophage polarization with increasing MD‐2 affinity.

Collectively, these findings demonstrate that engineering D‐peptide crosslinkers to enhance MD‐2 binding affinity amplifies the immunomodulatory capacity of MAP scaffolds. The concordance between docking predictions, SPR affinity measurements, MD2‐IN‐1 perturbation studies, and macrophage phenotypes supports MD‐2 engagement as a tunable design parameter for regulating macrophage activation. These results establish crosslinker residue chirality as a tunable and predictive design axis for controlling MAP–immune interactions, and they motivated subsequent in vivo studies to determine whether enhanced MD‐2 engagement translates into improved tissue‐level repair outcomes.

2.7. Enhanced MD‐2 Engagement via DMMP2 Preserves Regenerative Healing While Improving Tissue Remodeling Outcomes In Vivo

Given that DMMP2 exhibited enhanced MD‐2 engagement and stronger immunomodulatory effects in vitro, we next evaluated whether affinity‐guided peptide engineering preserved or improved the regenerative phenotype previously observed with DMAP in vivo. Importantly, the objective of these studies was not to directly validate the mechanism of action identified in vitro, but rather to determine whether engineering crosslinker chirality to enhance MD‐2 engagement could translate into comparable or improved tissue repair outcomes. Accordingly, DMAP, DMAP2, or sham controls were evaluated in a murine excisional wound model, and healing outcomes were assessed using wound closure measurements, histological scoring, and extracellular matrix analyses (Figure 7 and Figure S5).

FIGURE 7.

DMAP2 recapitulates the regenerative benefits of DMAP while reducing epidermal thickening and MAP degradation in vivo. (A) Representative H&E‐stained wound sections from control, DMAP‐, and DMAP2‐treated mice. Scale bars represent 1000 µm in the zoomed‐out images and 200 µm in the zoomed‐in images. (B) Composite histological scoring of re‐epithelialization, epidermal thickness index, keratinization, granulation tissue, remodeling, and scar elevation index. (C) Epidermal thickness ratios were quantified across treatment groups. DMAP2 demonstrated a significantly reduced epidermal thickness ratio relative to DMAP, indicating a more normalized epidermal architecture. Male and female analyses showed consistent trends. (D) Quantification of MAP material remaining within the wound bed. DMAP2‐treated wounds exhibited modestly increased MAP retention relative to DMAP, with sex‐stratified analyses showing significantly higher retention in males than females. (E) Follicle‐like metrics, including follicle‐like structures per millimeter and percentage of afollicular regions, were comparable between DMAP and DMAP2. Composite histological scores were analyzed using a mixed‐effects model with treatment matched within mouse, followed by Dunnett's multiple‐comparisons test. Overall DMAP–DMAP2 comparisons were performed using two‐tailed paired t‐tests. Sex‐stratified analyses were performed using two‐way mixed‐effects models with treatment as the within‐mouse factor and sex as the between‐mouse factor, followed by Fisher's LSD tests for planned within‐sex comparisons. p values are indicated above the bars. n = 5 female and 5 male mice.

Longitudinal macroscopic assessment demonstrated progressive wound closure in all treatment groups, with DMAP‐ and DMAP2‐treated wounds exhibiting similar healing trajectories and complete closure by day 15 (Figure S5A,B). No significant differences in overall wound closure rate were observed between DMAP and DMAP2, indicating that the tissue‐level differences identified at later time points were not attributable to accelerated wound closure.

Histological analysis at day 21, a time point corresponding to the remodeling phase of wound healing in this murine model [42], revealed that both DMAP‐ and DMAP2‐treated wounds exhibited robust regenerative phenotypes compared to sham controls, with composite histological scores approaching ∼10 out of a maximum score of 12, indicative of near‐complete regenerative healing (Figure 7A,B). Importantly, while overall histological scores were comparable between DMAP and DMAP2, qualitative inspection of H&E‐stained sections suggested a more normalized epidermal architecture in DMAP2‐treated wounds, with reduced epidermal thickening. Because DMAP already exhibited strong regenerative performance across multiple components of the composite histological score, including re‐epithelialization and restoration of follicle‐like structures approaching unwounded skin, the opportunity for further improvement in the overall composite score was inherently limited. We therefore focused on whether DMAP2 preserved these established regenerative features while further improving individual tissue‐level outcomes.

Quantitative analysis of epidermal thickness ratio, a metric reflecting re‐epithelialization quality and epidermal normalization, confirmed this observation. DMAP2‐treated wounds exhibited epidermal thickness ratios closer to 1, consistent with unwounded skin, and significantly lower than those observed in DMAP‐treated wounds, which showed ratios closer to 1.5, indicative of epidermal hypertrophy [42] (Figure 7C). Sex‐stratified analysis revealed that this normalization effect was most pronounced in male mice, while female cohorts displayed a similar but more modest trend toward reduced epidermal thickness with DMAP2 treatment.

We next assessed scaffold degradation within the wound bed as a measure of material stability and integration. DMAP2‐treated wounds retained modestly more MAP material than DMAP‐treated wounds at day 21, with sex‐stratified analyses revealing significantly greater material retention in males compared to females (Figure 7D). Increased scaffold persistence in DMAP2‐treated wounds may reflect reduced inflammatory cell‐mediated degradation, consistent with the stronger suppression of pro‐inflammatory macrophage activation and enhanced modulation of macrophage phenotype observed in vitro.

We evaluated follicular regeneration using follicle‐like metrics, including the number of follicle‐like structures per millimeter and the percentage of afollicular regions [8]. These measures were comparable between DMAP and DMAP2 treatment groups, indicating that both scaffolds similarly support hair follicle‐associated regenerative processes and that enhanced MD‐2 engagement did not compromise follicular regeneration (Figure 7E).

To further assess extracellular matrix remodeling, Picrosirius red staining was performed on wound sections collected at day 21 and compared to unwounded skin (Figure S5C,D). Under polarized light, both DMAP‐ and DMAP2‐treated wounds exhibited collagen fiber organization that closely resembled unwounded skin, with comparable fiber coherence and length. In contrast, collagen density and fiber width remained lower than those observed in unwounded tissue, indicating delayed matrix remodeling. Notably, no significant differences in collagen organization parameters were observed between DMAP and DMAP2, suggesting that enhanced MD‐2 engagement preserves normal extracellular matrix remodeling while maintaining the regenerative healing response.

Together, these in vivo findings demonstrate that affinity‐guided engineering of D‐chiral crosslinkers preserves the robust regenerative phenotype associated with DMAP while tuning selected tissue‐level outcomes, including epidermal normalization and scaffold persistence. These findings support the utility of peptide chirality as a tunable design parameter for optimizing MAP scaffold performance in regenerative wound healing.

3. Discussion

This study establishes peptide crosslinker chirality as a tunable regulator of immune responses within MAP scaffolds and identifies MD‐2‐associated TLR4 signaling as an important pathway linking molecular design to immune and tissue‐level outcomes. By integrating macrophage phenotyping, competitive inhibition, receptor‐binding analyses, pharmacologic perturbation studies, receptor trafficking measurements, reporter assays, temporal proteomics, and affinity‐guided peptide engineering, we demonstrate that D‐chiral peptide crosslinkers selectively reprogram macrophage activation through modulation of TLR4/MD‐2 signaling. Our findings further reveal that chirality‐dependent receptor engagement preserves receptor‐proximal innate immune sensing while attenuating downstream inflammatory signaling, ultimately promoting a more regenerative immune environment. Together, these results provide a mechanistic framework linking peptide‐level biomaterial design to immune regulation and tissue repair and establish peptide chirality as a design parameter for engineering immunomodulatory biomaterials.

We first demonstrate that MAP scaffolds broadly suppress inflammatory macrophage activation relative to 2D culture, but that crosslinker chirality exerts a dominant influence on the magnitude and quality of this response. Despite comparable physical properties between LMAP and DMAP scaffolds, including size, stiffness, porosity, and material composition [12], DMAP more effectively reduced classical M1 markers such as iNOS and CD86 and redistributed macrophage populations away from fully inflammatory states toward minimally activated M0‐like and transitional phenotypes. Importantly, this shift did not reflect global immune silencing, but rather a selective rebalancing of macrophage activation states, consistent with the concept that regenerative healing requires controlled modulation of inflammation [43, 44].

The requirement for scaffold‐bound presentation of D‐chiral peptides further underscores the importance of material context in immunomodulation. Soluble DMMP alone failed to recapitulate the effects of DMAP [11] yet competitively interfered with DMAP‐mediated macrophage modulation when added in excess, partially restoring inflammatory activation. This competitive behavior, which was markedly weaker for soluble LMMP, indicates that D‐chiral peptides directly engage macrophage receptors in a chirality‐dependent manner where scaffold‐associated presentation is required for effective signaling modulation. These findings suggest that MAP scaffolds act not simply as passive peptide depots, but as structured platforms that regulate receptor engagement dynamics.

While competitive inhibition studies, receptor binding analyses, and trafficking experiments strongly support direct receptor–peptide interactions as a key mechanism underlying DMAP‐mediated immunomodulation, these data do not exclude contributions from additional material‐dependent effects. The partial restoration of inflammatory signaling by soluble DMMP also indicates that multiple mechanisms contribute to the macrophage phenotypes induced by DMAP. In addition to direct receptor engagement, DMAP may alter the local microenvironment, including altered protein adsorption, cytokine sequestration [17, 45], or ligand presentation, which could influence both the magnitude and persistence of macrophage activation states.

Importantly, LMAP and DMAP are based on the same PEG‐VS MAP scaffold platform and exhibit comparable particle size, porosity, stiffness, adhesive ligand presentation, and overall material composition, with crosslinker chirality representing the primary design change [12]. Macrophages are highly responsive to physicochemical cues within biomaterials, including changes in stiffness, ligand presentation, spatial confinement, and surface properties [3, 14, 17]. Consistent with the absence of major differences in these bulk material characteristics, macrophages cultured in LMAP and DMAP displayed comparable average cell morphology despite exhibiting distinct activation trajectories (Figure S2). Together with the low‐fouling properties of PEG‐VS hydrogels [46], these observations reduce the likelihood that gross differences in material properties account for the observed biological responses and instead support a role for chirality‐dependent receptor interactions.

Moreover, DMAP reduced expression of antigen‐presenting markers such as CD11c and MHCII under M1‐polarizing conditions, effects that may reflect downstream macrophage silencing rather than TLR4‐MD2 signaling alone. This observation suggests that DMAP may initiate signaling through TLR4‐MD2 while also influencing broader macrophage polarization programs that contribute to the final phenotype. Although our findings support a role for TLR4‐MD2 signaling in the immunomodulatory effects of DMAP, they do not exclude contributions from additional receptor systems, downstream transcriptional programs, or scaffold‐mediated mechanisms. This represents an important limitation of the current study and motivates future investigations into the complementary pathways that contribute to macrophage reprogramming and tissue repair.

Through in silico docking and experimental validation, we identified MD‐2 as a preferential binding target for DMMP. Compared to LMMP, DMMP exhibited a higher predicted interaction confidence score, more extensive engagement within the MD‐2 lipid A binding pocket, and substantially higher binding affinity by SPR. Given MD‐2's essential role in LPS sensing and TLR4 activation [24], these results provide a plausible molecular basis for the observed suppression of M1 macrophage activation in DMAP scaffolds. Notably, DMMP binding did not exceed the affinity of LPS for MD‐2, suggesting that DMAP attenuates rather than abolishes TLR4 signaling, consistent with the balanced macrophage phenotypes observed.

The enhanced MD‐2 engagement observed with D‐chiral peptides likely arises from stereochemical changes introduced by D‐amino acid incorporation, as D‐chiral residues can alter crosslinker local backbone geometry, side‐chain orientation, and conformational flexibility relative to the corresponding all‐L crosslinker sequences [47]. To investigate these effects, we performed an in silico structural descriptor analysis of LMMP, DMMP, DMMP2, and additional stereochemical variants (Data S5). Interestingly, the peptide variants exhibited highly similar predicted net charge and hydrophobicity profiles despite displaying substantial differences in MD‐2 docking scores. These observations suggest that the enhanced MD‐2 engagement of DMMP and DMMP2 is unlikely to arise from large changes in bulk physicochemical properties alone.

Instead, the descriptor analysis and docking results are consistent with chirality‐dependent changes in peptide conformation and molecular presentation influencing receptor recognition. Consistent with this interpretation, DMMP and DMMP2 were predicted to engage a greater number of MD‐2 residues and occupy a larger portion of the MD‐2 lipid A binding pocket than LMMP. Furthermore, peptide variants containing additional D‐amino acid substitutions exhibited differences in backbone extension and overall molecular geometry, suggesting that chirality may influence both receptor accessibility and susceptibility to protease recognition through changes in peptide organization. While the current study was not designed to directly resolve the precise structural determinants governing MD‐2 recognition, our findings suggest that peptide chirality influences receptor binding through both stereochemical and conformational effects that can be leveraged as a tunable biomaterial design parameter. These chirality‐dependent changes in peptide conformation influence receptor recognition, raising the possibility that altered MD‐2 engagement may propagate to downstream signaling pathways.

To investigate the above hypothesis, we next examined how DMAP influences TLR4/MD‐2 trafficking and signaling dynamics. Beyond receptor binding, DMAP altered the intracellular and extracellular distribution of MTS510‐detectable TLR4/MD‐2 complexes and attenuated endocytosis‐dependent inflammatory signaling. Pharmacological inhibition of endocytosis phenocopied the effects of DMAP in LMAP‐treated macrophages but produced minimal additional effects in DMAP cultures, indicating that DMAP intrinsically modulates signaling pathways associated with TLR4 endocytosis. These observations were further supported by THP1‐Dual reporter studies demonstrating reduced NF‐κB and IRF activity and by temporal wound proteomics showing preservation of receptor‐proximal TLR4 sensing together with attenuation of downstream signaling and endocytosis‐dependent TRIF/IRF pathway components, including IFNB1. Collectively, these findings support a model in which DMAP preserves innate immune sensing while selectively rewiring downstream TLR4/MD‐2 signaling and trafficking.

Importantly, the TLR4–MD‐2 pathway is not exclusive to macrophages and plays critical roles in regulating the activation of other immune and stromal cell populations, including dendritic cells [37] and fibroblasts [48]. Thus, modifying peptide chirality within MAP scaffolds is likely to influence multiple cell types within the wound microenvironment, extending the immunomodulatory effects of DMAP beyond macrophages alone. This broader activity may help explain why, although DMAP promotes a more balanced macrophage phenotype [12], the overall regenerative response is characterized by an early increase in inflammatory and adaptive immune signaling, followed by later activation of keratinization, muscle regeneration, lipid formation, and additional skin repair pathways [8]. Future studies will be required to delineate how chirality‐dependent receptor engagement coordinates the responses of immune and connective tissue cells throughout the regenerative process.

This mechanistic insight enabled the rational design of next‐generation peptide crosslinkers. By engineering DMMP2, a D‐chiral peptide variant with enhanced MD‐2 affinity, we achieved graded amplification of immunomodulatory effects, including deeper suppression of pro‐inflammatory markers and increased expression of pro‐reparative markers in vitro. The strong correspondence between MD‐2 binding affinity and macrophage phenotype provides evidence that measured receptor affinity can inform the rational tuning of peptide design. Notably, DMMP2 contains only two D‐chiral residues, illustrating that subtle chirality changes can produce substantial functional outcomes in vitro.

The regenerative potential of this strategy was subsequently evaluated in vivo. In a murine excisional wound model, DMAP2 preserved the regenerative phenotype associated with DMAP while producing targeted improvements in selected tissue‐level outcomes, including epidermal normalization and scaffold persistence. Reduced epidermal thickening and increased material retention are consistent with, but do not directly demonstrate, a dampened inflammatory environment, as epidermal hypertrophy and accelerated scaffold degradation have both been associated with heightened inflammatory responses during wound healing [11, 49, 50]. Importantly, follicular regeneration was maintained, indicating that enhanced MD‐2 engagement does not compromise key regenerative processes of adnexal structures. The improved epidermal architecture and scaffold retention observed with DMAP2 are consistent with the enhanced MD‐2 binding and macrophage modulation demonstrated in vitro. Furthermore, temporal proteomic analyses of DMAP‐treated wounds revealed modulation of downstream TLR4‐associated signaling pathways, providing additional pathway‐level support for the proposed mechanism in vivo. However, additional studies employing genetic or pharmacological disruption of MD‐2 signaling in vivo will be required to directly test this relationship.

Together, these findings establish peptide crosslinker chirality as a tunable molecular design parameter for regulating immune responses within biomaterials. By integrating receptor binding analyses, pharmacologic perturbation studies, trafficking measurements, reporter assays, temporal proteomics, and affinity‐guided peptide engineering, we identify MD‐2‐associated TLR4 signaling as an important mechanistic link between crosslinker chirality and macrophage phenotype. Our data support a model in which D‐chiral peptides preserve receptor‐proximal innate immune sensing while selectively altering TLR4 trafficking and downstream inflammatory signaling, ultimately promoting a more regenerative immune environment. Importantly, enhanced MD‐2 engagement through rational peptide engineering preserved the regenerative benefits of DMAP and produced modest improvements in selected tissue‐level outcomes, demonstrating the potential of affinity‐guided biomaterial design. Collectively, this work positions MD‐2‐targeted peptide engineering as a strategy for tuning biomaterial–immune interactions and provides a framework for the development of next‐generation immunomodulatory MAP scaffolds.

4. Materials and Methods

4.1. Microparticle Generation and Purification

Microfluidic devices and microgels were produced as previously described [12, 13]. Briefly, 8‐arm Polyethylene Glycol‐Vinylsulfone (PEG‐VS, 20 kDa, JenKem) was dissolved in 0.3 M triethylamine (Sigma) pH 8.5 and pre‐reacted with K‐peptide (Ac‐FKGGERCG‐NH2, GenScript), Q‐peptide (Ac‐NQEQVSPLGGERCG‐NH2, GenScript), and RGD (Ac‐RGDSPGERCG‐NH2, GenScript) for at least one hour at 37°C. Then, the pH was adjusted to 4.5–5 to slow down the on‐chip gelation and prevent any clogging at the flow‐focusing region. The final stock precursor concentration was 10% (w/v) 8‐arm PEG‐VS with 1000 µM K‐peptide, 1000 µM Q‐peptide, and 1000 µM RGD. The cross‐linker solution was prepared by dissolving the di‐thiol matrix metalloproteinase‐sensitive peptide L‐MMP (Ac‐GCRDGPQGIWGQDRCG‐NH2, GenScript), D‐MMP (Ac‐GCRDGPQ_DGI_DW_DGQDRCG‐NH2, GenScript), or D‐MMP2 (Ac‐GCRDGPQGIW_DGQ_DDRCG‐NH2, in‐house‐made) in distilled water at 12 mM for 8‐arm LMAP or 12.4 mM for 8‐arm DMAP/DMAP2 and reacted with 10 µM Alexa Fluor 647‐maleimide (Thermo Fisher) for 5 min. These solutions were filtered through a 0.22 µm sterile filter, mixed 1:1, and loaded into a 1 mL Luer‐lock syringe. The pinching oil phase was a heavy mineral oil supplemented with 5% v/v Span‐80 (Sigma–Aldrich). The microgels were cross‐linked through pH‐mediated chemistry involving heavy mineral oil with 5% Span80 and Trimethylamine (2.5% v/v, Sigma–Aldrich) in the collection tube (2.5% v/v, Sigma–Aldrich). After gelation, the HMP and oil mixture was centrifuged to remove the oil‐phase top layer. The HMPs were then washed with 2% Pluronic F127 (Sigma Aldrich) in HEPES buffer (pH 8.3, 0.3 M) to assist with removal of the surfactant. Follow‐up washes were performed with HEPES buffer (pH 8.3, 0.3 M, 1% antibiotic–antibiotic (AA)) and centrifuged at 5000 RPM for 5 min. The purified microspheres were tested to ensure endotoxin levels were below threshold (Pierce) and then stored in HEPES buffer (pH 8.3, 0.3 M, 1% AA, 10 mM CaCl₂) at 4°C.

4.2. BMDM Isolation, Differentiation, and Culture in MAP Scaffolds

Bone marrow was harvested from femurs and tibias of 8–12 week‐old C57BL/6 mice (mixed sex) [13]. Cells were cultured in non‐tissue‐culture‐treated flasks in IMDM + 10% FBS + 1% Anti‐Anti supplemented with 15 ng/mL M‐CSF (PeproTech). Media was refreshed every 3 days, and macrophages were harvested on day 10 post‐extraction using 1X TrypLE with 10 mM EDTA for detachment. For embedding into scaffolds, BMDMs were suspended at 10 000 cells/µL, pelleted down, and purified HMPs previously swelled in media were pipetted on top of the cell pellet, using a displacement pipette. 1 µL thrombin (200 U/mL in 200 mM Tris‐HCl, 150 mM NaCl, 20 mM CaCl₂) and 2 µL Factor XIII (250 U/mL) were added on top of the HMPs, and the whole mixture was mixed and allowed to anneal in individual 50 µL gels for 45 min on top of Sigma‐coat slides. After annealing, gels were submerged in media in a six‐well plate, and parallel 1 × 10⁶ cells/well were seeded as 2D controls. After 24 h, cells were stimulated with LPS (20 ng/mL) + IFN‐γ (20 ng/mL) for 24 additional hours to trigger M1 activation. To test peptide–receptor competition, soluble DMMP or LMMP (24 mM) was added to BMDMs cultured in DMAP scaffolds at the time of polarization. Cells were cultured under M1 conditions for 24 h before flow cytometry, imaging, or ELISA.

4.3. Flow Cytometry Staining, Gating, and Quantification

Cells were retrieved from scaffolds using a combination of enzymatic and mechanical dissociation. For scaffold‐based cultures, samples were incubated in 1× TrypLE supplemented with 10 mM EDTA, followed by mechanical disruption, and then incubation at 37°C for 10 min. Isolated cells were washed, pelleted, and stained for viability using Zombie NIR (BioLegend) for 15 min at room temperature. Cells were then blocked with FcR Blocking Buffer (Miltenyi Biotec) for 10 min at 4°C. Staining was performed by incubating cells with antibodies against CD11b and F4/80 (macrophage markers), CD86 and iNOS (intra‐ and extracellular pro‐inflammatory markers), MHCII and CD11c (extracellular antigen‐presenting markers), and CD206 and Arg1 (intra‐ and extracellular reparative markers), and TLR4/MD‐2 (intra‐ and extracellular) (Tables S2 and S3) for 30 min at 4°C. For intracellular staining, the cells were subsequently fixed and permeabilized using the Foxp3/Transcription Factor Staining Buffer Set (eBioscience) according to the manufacturer's instructions, washed, and incubated with antibodies against intracellular markers, including intracellular TLR4/MD‐2, iNOS, Arg‐1 for 30 min at 4°C. For the TLR4/MD‐2 intra/extracellular ratio, extracellular staining occurred before permeabilization, and intracellular staining occurred after fixation/permeabilization. The TLR4/MD‐2 antibody used in this study, clone MTS510, recognizes a conformation‐sensitive epitope on the TLR4/MD‐2 complex whose accessibility can change following receptor activation. Ratios were calculated as

$$\mathrm{Intracellular}/\mathrm{Extracellular}\mathrm{Ratio}=\frac{\mathrm{MF}{\mathrm{I}}_{\mathrm{intracellular}}}{\mathrm{MF}{\mathrm{I}}_{\mathrm{surface}}}$$

4.4. Pharmacologic Inhibition of MD‐2 Signaling

To evaluate the contribution of MD‐2 signaling to macrophage activation, BMDMs were cultured on 2D tissue culture plastic or within LMAP and DMAP scaffolds as described above. 24 h after seeding, cultures designated for MD‐2 inhibition were treated with the MD‐2 inhibitor MD2‐IN‐1 (TargetMol) at a final concentration of 10 µM and incubated for 1 h prior to M1 polarization. Macrophages were subsequently treated with LPS and IFNγ using the same conditions described for M1 activation experiments in Section 4.2. Following 24 h of stimulation, cells were harvested and analyzed by flow cytometry using the same staining panel and gating strategy described above. Untreated cultures that did not receive MD2‐IN‐1 served as controls.

4.5. Dynasore‐Mediated Endocytosis Inhibition

Endocytosis inhibition of the TLR4/MD‐2 complex using dynasore was performed as previously described [34]. Dynasore (Sigma–Aldrich) was used at a final concentration of 80 µM to inhibit endocytosis. BMDMs were pretreated with dynasore for 60 min prior to M1 stimulation and subsequently cultured for 12 or 24 h under 2D, LMAP, or DMAP conditions. Flow cytometry was used to quantify the marker panel described in Tables S2 and S3. To assess the contribution of sustained versus transient endocytosis inhibition, dynasore was either maintained throughout the duration of M1 stimulation or removed after the initial 1‐hour pretreatment period.

4.6. ELISA Quantification of TNF Secretion

Supernatants were collected after 24 h of BMDM polarization prior to BMDM extraction from MAP scaffolds and analyzed using a mouse TNF ELISA kit (Thermo Scientific) per manufacturer instructions. Absorbance was detected at 450 nm with a Tecan plate reader, and concentrations were determined from standard curves.

4.7. High‐Dimensional Phenotyping and sUMAP Analysis

Flow cytometry data (.fcs files) were analyzed using FlowJo v10 (BD Biosciences). FlowClean was applied to remove acquisition anomalies [51], and datasets were down‐sampled to an equal number of events per sample using the DownSample plugin to minimize sampling bias. Dimensionality reduction was first performed using unsupervised UMAP [52] to explore cellular heterogeneity and inform the appropriate number of subclusters. Based on these results, unsupervised clustering was then performed using the FlowSOM plugin [53] with the number of clusters set accordingly. FlowSOM‐derived cluster assignments were subsequently used to generate a sUMAP projection for final dimensionality reduction and visualization of phenotypic relationships among cells.

4.8. THP‐1 Dual Reporter Assay

To assess whether chirality‐dependent modulation of innate immune signaling extends to a human macrophage‐like system, THP1‐Dual reporter cells (InvivoGen) were used. THP1‐Dual cells enable simultaneous monitoring of NF‐κB activity through secretion of SEAP and IRF activity through secretion of luciferase. THP1‐Dual cells were maintained according to the manufacturer's recommendations. For reporter experiments, cells were cultured in RPMI supplemented with 10% HI‐FBS and 1% AA. Cells were seeded at 5 × 10⁵ cells per well in six‐well plates for untreated 2D negative controls and LPS‐stimulated 2D controls, or embedded at the same density within 50 µL LMAP or DMAP scaffolds. Monocytes were differentiated into macrophage‐like cells by treatment with PMA (20 ng/mL) for 3 h, followed by a 72‐hour recovery period. Following differentiation, cells were stimulated with LPS (10 ng/mL), and culture supernatants were collected at 12 and 24 h post‐stimulation. NF‐κB and IRF reporter activities were quantified using QUANTI‐Blue and QUANTI‐Luc assays, respectively.

4.9. Peptide Design and In Silico Docking (HDOCK)

To enable in silico docking of LMMP and nonnatural DMMP‐containing peptides incorporating D‐amino acid chirality, peptide structures were generated using PEPstrMOD [20]. The resulting peptide models were exported as PDB files and used as ligands for docking analyses. Protein–peptide docking was performed using the HDOCK server [18], which supports protein–protein and protein–peptide docking with PDB‐formatted input structures. Each peptide was docked against the corresponding protein of interest, and docking poses were ranked based on the HDOCK Confidence Score.

For downstream analysis and visualization, the protein–peptide complex with the highest confidence score for each pairing was selected and presented in Figures 3 and 6. Confidence Score values were interpreted as follows: scores > 0.7 indicate a high likelihood of binding, scores between 0.5 and 0.7 suggest possible binding, and scores < 0.5 indicate a low likelihood of binding. Because the confidence score is derived from an empirical scoring function, it was interpreted cautiously and used primarily for comparative ranking rather than absolute binding affinity prediction [18].

4.10. Surface Plasmon Resonance

SPR was performed on a Biacore T200. Recombinant mouse MD‐2/LY96 protein was purchased from MedChem Express (Cat. No. HY‐P77081) and immobilized on a CM5 sensor chip using amine coupling after EDC/NHS activation (pH 5.0 sodium acetate buffer) to ∼1200 RU. LMMP, DMMP, DMMP2 (3–6 mM), and LPS (positive control) were injected in HBS‐EP+ buffer at 30 µL/min. Surfaces were regenerated with 10 mM glycine pH 2.0. Because LMMP and DMMP exhibited rapid association and dissociation kinetics that precluded robust kinetic parameter estimation, equilibrium binding responses were analyzed using steady‐state fitting to estimate K _D values. This approach is commonly used for low‐affinity interactions that do not produce well‐resolved association and dissociation phases [54].

4.11. Peptide Synthesis

The DMMP2 peptide (Ac‐GCRDGPQGIW_DGQ_DDRCG) was synthesized via standard Fmoc solid‐phase peptide synthesis (SPPS) on a CEM Liberty Blue peptide synthesizer using Rink amide resin. Following synthesis, the peptide was cleaved from the resin using a cocktail of 92.5% trifluoroacetic acid (TFA), 2.5% triisopropylsilane, 2.5% 1,2‐ethanedithiol, and 2.5% deionized (DI) water for 2 h at room temperature with constant agitation. The cleavage solution was concentrated by rotary evaporation to remove excess TFA, followed by peptide precipitation and multiple washes with cold diethyl ether to remove residual scavengers and cleavage reagents. The crude peptide was then purified by reverse‐phase HPLC using a Waters XBridge BEH C18 column with a water–acetonitrile mobile phase. Peptide identity (m/z) was confirmed using a Bruker MALDI‐TOF mass spectrometer prior to lyophilization and storage at −20°C until use.

4.12. Peptide Structure Preparation and In Silico Descriptor Analysis

Three‐dimensional peptide structures for LMMP and D‐amino acid‐containing peptide variants were generated using PEPstrMOD and exported in PDB format. Structures were cleaned using a custom Python script to retain only valid coordinate records prior to downstream analysis. Cleaned PDB files were converted to MOL2 format using Open Babel (v3.x) with explicit hydrogen addition. MOL2 files were parsed in Python to calculate peptide composition and charge‐related descriptors. Molecular weight, elemental composition, estimated net charge, residue‐level charge distribution, and counts of strongly charged functional groups were calculated from atom identities and partial charges contained within the MOL2 files. These values were compiled in the Physicochemical_Descriptors csv of Data S5.

To compare structural features among peptide variants, cleaned PDB models were further analyzed using custom Python scripts. The radius of gyration calculated using all atoms (Radius_of_Gyration_AllAtoms), backbone radius of gyration calculated using Cα atoms (Radius_of_Gyration_Backbone), and N‐to‐C terminal distances (N_to_C_Terminal_Distance) were calculated from atomic coordinates as measures of peptide compactness and backbone extension. Sequence hydrophobicity descriptors were determined using the Kyte–Doolittle hydrophobicity scale, including the hydrophobic residue fraction (Hydrophobic_Residue_Fraction) and mean hydrophobicity index (Mean_Kyte_Doolittle_Index). These descriptors were compiled in the Structural_Geometry worksheet of Data S5.

Descriptor values were used to compare LMMP, DMMP, DMMP2, and additional stereochemical variants with respect to charge distribution, hydrophobicity, molecular geometry, and overall structural organization. A summary of peptide composition and charge descriptors is provided in the Peptide_Summary worksheet of Data S5. Custom Python scripts used for structure cleaning, descriptor extraction, and geometric analyses are provided in Data S6.

4.13. In Vivo Wound Healing Study

Male and female SKH‐1 Elite mice (8–12 weeks old) were obtained from Charles River Laboratories and housed in a centralized animal facility at Duke University. All animal procedures were conducted in accordance with an Institutional Animal Care and Use Committee (IACUC)‐approved protocol (A001‐24‐01). An excisional splinted wound model was used as previously described by us and others [8]. Briefly, mice were anesthetized with 4% isoflurane for induction and maintained at 1.5%–2% isoflurane on a heating pad throughout the procedure. Sustained‐release buprenorphine (0.5 mg/kg; 0.5 mg/mL) was administered subcutaneously for analgesia. The dorsal surface was shaved and prepared using alternating scrubs of iodine and 70% isopropyl alcohol. Two full‐thickness excisional wounds were created on each side of the dorsal midline using a sterile 5 mm biopsy punch, yielding four total wounds per animal.

Adherent polydimethylsiloxane (aPDMS) ring splints with a 7 mm internal diameter were affixed circumferentially around each wound to minimize contraction and promote healing via re‐epithelialization and granulation tissue formation [55]. HMPs were prepared on the day of surgery by centrifugation and removal of supernatant, then mixed with thrombin (final concentration 4 U/mL; stock 200 U/mL in 200 mM Tris‐HCl, 150 mM NaCl, 20 mM CaCl₂) and Factor XIII (final concentration 10 U/mL; stock 250 U/mL). The resulting mixture was immediately injected into wounds (10 µL per wound) to allow in situ gelation of the MAP scaffolds.

Each mouse received a randomized distribution of treatments across wounds, including DMAP2, DMAP, and sham controls, with a fourth wound assigned to an exploratory condition not included in the present study. Each treatment was represented once per animal, and the experimental unit (N) was defined as an individual mouse. Following material administration and a 30 min gelation period, wounds were covered with Tegaderm dressings to seal the splints, prevent desiccation, and protect against mechanical disruption.

Animals were housed individually with environmental enrichment and monitored daily for body weight, splint integrity, and dressing adherence. Digital photographs of wounds were collected immediately following surgery (day 0), at day 4, day 7, day 10, and at day 15 post‐wounding using a Handheld Digital Microscope Pro (Celestron) to assess wound progression and calculate percent wound closure. Complete wound images, wound closure measurements, body weight records, and wound‐site assignments for each animal are provided in Data S7. On day 21 post‐wounding, mice were euthanized, and wound tissues were excised for downstream histological and extracellular matrix analyses.

4.14. Sample Collection

Mice were euthanized, and dorsal skin was carefully excised. On day 21 post‐wounding, 8 mm biopsy punch samples encompassing the wound sites were collected, bisected, and fixed in 4% paraformaldehyde for 4 h at room temperature, followed by overnight fixation. Samples were then transferred to 70% ethanol prior to paraffin embedding. Paraffin‐embedded tissues were sectioned at 5 µm thickness for histological staining.

4.15. Histological Tissue Staining

For all histological stains, paraffin sections were first deparaffinized and rehydrated through graded xylene and ethanol solutions. Hematoxylin and eosin (H&E) staining was performed by incubating sections in Mayer's hematoxylin solution (EMS) for 15 min, followed by bluing in warm tap water for 15 min. Sections were then rinsed in DI water and 95% ethanol before counterstaining with alcoholic eosin Y (EMS) for 40 s. Following completion of staining, sections were dehydrated and cleared through graded ethanol and xylene, then mounted with Permount (Fisher Scientific). Brightfield images were acquired using a Zeiss Axio Scan at 20× magnification.

4.16. Histological Analysis

All tissue analyses were performed in a blinded manner. The number of secondary appendages within the wound bed was quantified and normalized to the total wound length. The afollicular percentage was calculated as the maximum distance between adjacent dermal follicular appendages divided by the total wound length, yielding the proportion of the wound bed lacking hair follicles, as described previously [8]. Histological evaluation was performed using a scoring system adapted from the previously described SPOT framework, incorporating assessments of scar elevation index, tissue remodeling, granulation tissue formation, keratinization, epidermal thickness index, and re‐epithelialization [42].

4.17. Picrosirius Red Staining and Collagen Fiber Analysis

Paraffin‐embedded wound sections (5 µm) were deparaffinized through sequential washes in xylene and graded ethanol solutions, followed by rehydration in distilled water. Sections were stained with Picrosirius red solution for 1 h, washed twice in acidified water, dehydrated through graded ethanol and xylene, and mounted using Permount mounting medium. Collagen organization was assessed under polarized light microscopy using a Zeiss Axio Observer 7 inverted microscope equipped with polarized light optics. Images were acquired at 40× optical magnification. For each wound, three nonoverlapping regions of interest (ROIs; 150 × 150 µm) were selected within the wound bed and analyzed in a blinded manner. Image analysis was performed using ImageJ/Fiji. Collagen fibers were segmented using the IsoData thresholding algorithm to generate binary masks for fiber identification. These masks were used to quantify collagen fiber density, fiber length, and fiber width. Collagen fiber alignment was assessed using the OrientationJ plugin, and coherence values were calculated as a measure of collagen organization and directional uniformity. Measurements from the three ROIs were averaged to generate a single value for each wound.

4.18. Targeted Analysis of Previously Published Temporal Wound Proteomics Data

To evaluate whether the in vitro observations of altered TLR4/MD‐2 signaling were reflected in vivo and to investigate the temporal dynamics of innate immune signaling during wound healing, a targeted analysis was performed using a previously published wound proteomics dataset comparing LMAP‐ and DMAP‐treated wounds at days 4, 14, and 21 post‐injury. The complete proteomics workflow, sample preparation, mass spectrometry acquisition, and data processing have been previously described [8], and the dataset is publicly available through the PRIDE repository (accession PXD067152). A subset of proteins associated with TLR4 signaling, innate immune sensing, NF‐κB/MAPK signaling, ROS/NADPH oxidase activity, and TRIF‐IRF‐type I interferon signaling was identified based on literature curation and pathway annotations. Approximately 90 proteins were selected for downstream analysis. Protein abundance values were extracted from the published dataset and analyzed using a pathway‐focused approach. Temporal expression patterns across LMAP and DMAP treatment groups were visualized using heatmaps generated with Heatmapper [56]. Differential protein abundance between DMAP‐ and LMAP‐treated wounds at day 4 was assessed using volcano plots generated with VolcaNoseR [57].

4.19. Biological Replicates and Statistics

For in vivo studies, wounds were randomized across treatment groups within each mouse, with each mouse receiving one wound per treatment condition. Therefore, treatment measurements obtained from wounds within the same mouse were analyzed as matched observations. Comparisons between two treatment conditions were performed using two‐tailed paired t‐tests. Endpoints involving three treatment conditions were analyzed using mixed‐effects models with mouse treated as the matched factor. Sex‐stratified analyses were performed using two‐way mixed‐effects models with treatment as the within‐mouse factor and sex as the between‐mouse factor. Each plotted wound‐level measurement represents a unique mouse‐treatment combination, and the biological replicate (n) corresponds to an individual mouse. All in vitro experiments were performed in triplicate using independent biological replicates, with BMDMs isolated from different mice and THP1‐Dual reporter assays performed using independent cell cultures. Statistical analyses were conducted using GraphPad Prism (version 10; GraphPad Software). One‐way or two‐way ANOVA, as specified in the legend for each experiment, followed by Tukey's post hoc test, was used to compare treatment groups. For targeted reanalysis of the published proteomics dataset, statistical comparisons between LMAP‐ and DMAP‐treated wounds were performed using two‐tailed Welch's t‐tests. Differences were considered statistically significant at p < 0.05. Unless otherwise indicated, data are presented as mean ± standard error of the mean (S.E.M.) No data transformations were applied unless stated otherwise. No outliers were excluded from statistical analyses, and all statistical tests were two‐sided.

Author Contributions

ASA and JT made substantial contributions to the study design, as well as to data acquisition, analysis, and interpretation. YL conducted experiments that contributed substantially to Figure 1. PCA performed experiments that contributed substantially to Figure 5 and Figure S1. AK performed experiments that contributed substantially to Figure 6. AE performed experiments that contributed substantially to Figures 2 and 3. LH and KP carried out in silico experiments that contributed substantially to Figure 3. FL performed image analysis that contributed substantially to Figure S5. ZY performed image analysis that contributed substantially to Figure S1. ASA and JT drafted the manuscript, and all authors discussed the results and contributed to writing and editing the manuscript. TS provided guidance and discussion throughout the project and made substantial contributions to experimental design, data analysis, and manuscript editing.

Funding

This research was supported by the National Institutes of Health under grants R01AI152568. J.L.T. was supported by the Duke University Center for Biomolecular and Tissue Engineering through NIH training grant T32GM144291. Surface plasmon resonance (SPR) experiments were supported by the 2024‐2025 Duke University Shared Materials Instrumentation Facility (SMIF) Voucher Program.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File 1: adhm71572‐sup‐0001‐suppdataS1‐S7.zip.

Supporting File 2: adhm71572‐sup‐0002‐figureS1‐S5.zip.

Data Availability Statement

All data supporting the findings of this study, including flow cytometry FCS files, ELISA quantifications, SPR sensorgrams, and docking input/output files, are available in The Research Data Repository at Duke University at https://doi.org/10.7924/r4r537. The temporal wound proteomics dataset reanalyzed in this study was previously published and is publicly available through the PRIDE repository under accession number PXD067152 and the Research Data Repository at Duke University at https://doi. org/10.7924/r4q245n68. Additional data related to this study are available from the corresponding author upon reasonable request.