Mechanical insights into regulation of bio–nano interactions for lipid-based nanocarriers Authors: Fan Yang, Yikai Wang, Yuhan Lu, Lin Lin, Yiyang Wang, Xiaochun Sun, Meijuan Liu, Ciro Chiappini, Yi-Feng Wang, Xing-Jie Liang Published: 2026 ABSTRACT Lipid nanocarriers (LNCs), including liposomes, lipid nanoparticles (LNPs), and extracellular vesicles (EVs), are promising platforms for targeted drug delivery. However, their clinical translation is hindered by multi-level heterogeneity and complex behaviors at the bio–nano interface. Interest in mechanotargeted delivery strategies has grown rapidly, as the mechanical properties of LNCs are increasingly recognized as critical determinants of their in vivo behavior. Establishing a comprehensive database of mechanical properties to decipher the underlying regulatory mechanisms is a prerequisite for the active modulation of LNCs’ mechanics and the rational design of delivery vehicles. Although studies on LNC mechanics are increasing, several challenges remain, including mechanical heterogeneity, unbalanced research across different carriers, low-throughput mechanical characterization and limited comparability among results from different laboratories. This review systematically summarizes the intrinsic mechanical parameters of the three types of LNCs, outlines the structural determinants of these properties, highlights their roles as key modulators at the bio–nano interface, and discusses advanced engineering strategies for mechanical modulation. The goal of this review is to facilitate mechanics-informed design of next-generation nanomedicines for precision therapy. Introduction Targeted delivery of therapeutic molecules using lipid-based nanocarriers (LNCs) has emerged as a transformative approach across oncology, infectious disease, and genetic medicine [1, 2]. The clinical development of this field began with the approval of Doxil in 1995. This PEGylated liposomal formulation of doxorubicin prolonged circulation, altered tissue distribution, and reduced cardiotoxicity compared with free drug [3]. AmBisome followed a similar principle, where the liposomal incorporation of amphotericin B changed the drug’s distribution and release behavior, reducing nephrotoxicity at therapeutic doses [4]. Beyond liposomes, ionizable lipid nanoparticles (LNPs) broadened nanocarrier applications to nucleic acid therapeutics. Patisiran (Onpattro) demonstrated durable hepatic TTR silencing through repeated systemic siRNA delivery [5]. The subsequent success of mRNA-LNP vaccines against COVID-19 validated this platform for rapid clinical deployment at a population scale [6]. Alongside these synthetic systems, extracellular vesicles (EVs) are naturally secreted lipid-bilayer nanoparticles (NPs) that mediate intercellular communication, which are now being evaluated as inherently biocompatible nanocarriers, with ongoing trials spanning osteoarthritis and gene therapy [7]. Together, liposomes, LNPs, and EVs represent the three major LNC platforms. The physicochemical properties of LNC significantly affect their delivery process and biological effects. With growing insights into bio–nano interactions, it is increasingly evident that mechanical properties such as stiffness, bending rigidity, and membrane phase state can influence how NPs interact with biological environments during transport, cellular entry, and intracellular trafficking [8–10]. When NPs enter biological fluids, proteins and other biomolecules adsorb onto their surface and form a biomolecular corona, which can reshape the bio–nano recognition, uptake route, and downstream trafficking [11, 12]. The mechanical state of the NP influences which proteins are recruited to the corona [13]. Intracellularly, the molecular geometry of ionizable lipids induces phase transitions that disrupt endosomal membranes, enabling efficient escape [14], whereas cholesterol-enriched EVs bypass endosomal confinement by increasing membrane fusion competence for direct cytosolic entry [15]. Beyond the NP itself, the mechanical state of the target cell also regulates delivery outcomes, since plasma membrane tension modulates membrane invagination and endocytosis, which can be modulated with external mechanical cues [16–20]. Together, these observations indicate that mechanical forces, whether intrinsic to the NP, governed by cellular state, or externally applied, constitute a distinct and underexplored axis in delivery design. Despite this progress, the mechanical properties of LNCs remain comparatively underexplored relative to their composition, surface chemistry, and cargo design. This review mainly focuses on research progress in LNCs over the past 5–10 years, centering on the correlation between LNC mechanics and biological effects across three categories. The first category addresses quantitative mechanical characterization of EVs, liposomes, and LNPs using terms such as bending rigidity, stiffness, Young’s modulus, membrane fluidity, and techniques including atomic force microscopy (AFM). The second category links mechanical properties to bio–nano interfacial processes by combining mechanical descriptors with protein corona formation, membrane wrapping, and endocytosis. The third category focuses on deliberate mechanical engineering strategies by searching for studies reporting tunable stiffness or deformability in the context of therapeutic performance. Studies were included when they reported quantitative mechanical data or established a link between mechanical properties and biological behavior, whether directly demonstrated or mechanistically inferred. LNC mechanics: crucial modulators of bio–nano interaction Conceptual definition: from chemical composition to physical entities Biophysically, LNCs behave as viscoelastic soft matter governed by distinct molecular architectures and are heterogeneous in stiffness, membrane bending rigidity, and membrane fluidity. Consequently, redefining LNCs as mechanical entities elucidates the physical forces driving their in vivo behavior, providing a novel engineering dimension for targeted delivery. Stiffness Stiffness characterizes the ability of an intact nanocarrier to resist macroscopic deformation under external mechanical stress. It is quantified using two parameters, stiffness (K, N/m) and Young’s modulus (E, Pa) [28, 29]. The former is an extrinsic structural property dictated by the size and geometry of the LNCs, whereas the latter is an intrinsic material property, independent of vesicle geometry, that defines the uniaxial stress–strain relationship. The stiffness of liposomes varies widely across formulations (Table 1). Pure DPPC vesicles exhibit a Young’s modulus of 110 ± 15 MPa [21]. In terms of stiffness K, DSPC-based mixtures range from 4.0 × 10−2 to 9.3 × 10−2 N/m, peaking at a 90:10 molar ratio of DSPC to DSTAP [22]. Conversely, DOPC-based liposomes are generally softer. The stiffness of pure DOPC liposomes is 3.3 × 10−2 N/m, while its 50:50 mixtures with DOTAP or DOPG drop to 2.2 × 10−2 and 1.9 × 10−2 N/m, respectively. Lastly, 50:50 cholesterol formulations yield 2.4 × 10−2 N/m (DPPC/CHOL) and 1.6 × 10−2 N/m (EPC/CHOL) [23].Table 1The mechanical properties of liposomesMechanical propertiesLiposomesValueReferencesStiffnessDPPC (100)110 ± 15 MPa[21]DSPC/DSTAP (90/10)DSPC/DSTAP (50/50)DSPC/DSPG (50/50)DOPC (100)DOPC/DOTAP (50/50)DOPC/DOPG (50/50)9.3 × 10–2 N/m5.2 × 10–2 N/m4.0 × 10–2 N/m3.3 × 10–2 N/m2.2 × 10–2 N/m1.9 × 10–2 N/m[22]DPPC/CHOL (50/50)EPC/CHOL (50/50)2.4 × 10–2 N/m1.6 × 10–2 N/m[23]Bending rigidityLiquid DisorderedLiquid OrderedSolid Ordered10–19 Jseveral 10–19 J10–18 J[24]POPC/CHOL (50/50)EPC/CHOL (50/50)0.83 × 10–19 J0.92 × 10–19 J[23]Membrane fluidityEPCEPC/CHOL (50/50)DSPC/CHOL (10/50)DSPC/CHOL (50/50)TMA-DPH r0.1950.2650.2880.187[25]SM-102/DSPC/CHOL/DMG-PEG2k6Ac1-C12/18PA/CHOL/DMG-PEG2k6Ac1-C12/DOPE/CHOL/DMG-PEG2k6Ac1-C12/DOPC/CHOL/DMG-PEG2k6Ac1-C12/POPC/CHOL/DMG-PEG2k6Ac1-C12/DSPC/CHOL/DMG-PEG2k6Ac1-C12/DPPC/CHOL/DMG-PEG2k6Ac1-C12/DMPC/CHOL/DMG-PEG2kLaurdan Probe GP0.470.290.420.400.430.500.470.48[26]DSPC/CHOL/DSPE-PEG200037 °C + Dextrose + 20 mol% CHOL37 °C + Glycerin + 20 mol% CHOL37 °C + Dextrose + 40 mol% CHOL37 °C + Glycerin + 40 mol% CHOL45 °C + Dextrose + 20 mol% CHOL45 °C + Glycerin + 20 mol% CHOL45 °C + Dextrose + 40 mol% CHOL45 °C + Glycerin + 40 mol% CHOLLaurdan Probe GP−13.263.353.263.483.623.713.864.20[27]References [21–24] all performed in liquid environments, but employed different substrates: silicon (Ref. [21]), BSA-coated glass (Refs. [22] and [24]), and APTES-treated mica (Ref. [23]). Regarding the mechanical parameters, Refs. [22] and [23] used the slope of the linear region of the force–deformation curve as the stiffness, Ref. [21] extracted the Young’s modulus based on Shell theory, while Refs. [23] and [24] further quantified the bending modulus (both based on Shell theory) Table 1The mechanical properties of liposomesMechanical propertiesLiposomesValueReferencesStiffnessDPPC (100)110 ± 15 MPa[21]DSPC/DSTAP (90/10)DSPC/DSTAP (50/50)DSPC/DSPG (50/50)DOPC (100)DOPC/DOTAP (50/50)DOPC/DOPG (50/50)9.3 × 10–2 N/m5.2 × 10–2 N/m4.0 × 10–2 N/m3.3 × 10–2 N/m2.2 × 10–2 N/m1.9 × 10–2 N/m[22]DPPC/CHOL (50/50)EPC/CHOL (50/50)2.4 × 10–2 N/m1.6 × 10–2 N/m[23]Bending rigidityLiquid DisorderedLiquid OrderedSolid Ordered10–19 Jseveral 10–19 J10–18 J[24]POPC/CHOL (50/50)EPC/CHOL (50/50)0.83 × 10–19 J0.92 × 10–19 J[23]Membrane fluidityEPCEPC/CHOL (50/50)DSPC/CHOL (10/50)DSPC/CHOL (50/50)TMA-DPH r0.1950.2650.2880.187[25]SM-102/DSPC/CHOL/DMG-PEG2k6Ac1-C12/18PA/CHOL/DMG-PEG2k6Ac1-C12/DOPE/CHOL/DMG-PEG2k6Ac1-C12/DOPC/CHOL/DMG-PEG2k6Ac1-C12/POPC/CHOL/DMG-PEG2k6Ac1-C12/DSPC/CHOL/DMG-PEG2k6Ac1-C12/DPPC/CHOL/DMG-PEG2k6Ac1-C12/DMPC/CHOL/DMG-PEG2kLaurdan Probe GP0.470.290.420.400.430.500.470.48[26]DSPC/CHOL/DSPE-PEG200037 °C + Dextrose + 20 mol% CHOL37 °C + Glycerin + 20 mol% CHOL37 °C + Dextrose + 40 mol% CHOL37 °C + Glycerin + 40 mol% CHOL45 °C + Dextrose + 20 mol% CHOL45 °C + Glycerin + 20 mol% CHOL45 °C + Dextrose + 40 mol% CHOL45 °C + Glycerin + 40 mol% CHOLLaurdan Probe GP−13.263.353.263.483.623.713.864.20[27]References [21–24] all performed in liquid environments, but employed different substrates: silicon (Ref. [21]), BSA-coated glass (Refs. [22] and [24]), and APTES-treated mica (Ref. [23]). Regarding the mechanical parameters, Refs. [22] and [23] used the slope of the linear region of the force–deformation curve as the stiffness, Ref. [21] extracted the Young’s modulus based on Shell theory, while Refs. [23] and [24] further quantified the bending modulus (both based on Shell theory) The mechanical properties of liposomes DSPC/DSTAP (90/10) DSPC/DSTAP (50/50) DSPC/DSPG (50/50) DOPC (100) DOPC/DOTAP (50/50) DOPC/DOPG (50/50) 9.3 × 10–2 N/m 5.2 × 10–2 N/m 4.0 × 10–2 N/m 3.3 × 10–2 N/m 2.2 × 10–2 N/m 1.9 × 10–2 N/m DPPC/CHOL (50/50) EPC/CHOL (50/50) 2.4 × 10–2 N/m 1.6 × 10–2 N/m Liquid Disordered Liquid Ordered Solid Ordered 10–19 J several 10–19 J 10–18 J POPC/CHOL (50/50) EPC/CHOL (50/50) 0.83 × 10–19 J 0.92 × 10–19 J EPC EPC/CHOL (50/50) DSPC/CHOL (10/50) DSPC/CHOL (50/50) TMA-DPH r 0.195 0.265 0.288 0.187 SM-102/DSPC/CHOL/DMG-PEG2k 6Ac1-C12/18PA/CHOL/DMG-PEG2k 6Ac1-C12/DOPE/CHOL/DMG-PEG2k 6Ac1-C12/DOPC/CHOL/DMG-PEG2k 6Ac1-C12/POPC/CHOL/DMG-PEG2k 6Ac1-C12/DSPC/CHOL/DMG-PEG2k 6Ac1-C12/DPPC/CHOL/DMG-PEG2k 6Ac1-C12/DMPC/CHOL/DMG-PEG2k Laurdan Probe GP 0.47 0.29 0.42 0.40 0.43 0.50 0.47 0.48 DSPC/CHOL/DSPE-PEG2000 37 °C + Dextrose + 20 mol% CHOL 37 °C + Glycerin + 20 mol% CHOL 37 °C + Dextrose + 40 mol% CHOL 37 °C + Glycerin + 40 mol% CHOL 45 °C + Dextrose + 20 mol% CHOL 45 °C + Glycerin + 20 mol% CHOL 45 °C + Dextrose + 40 mol% CHOL 45 °C + Glycerin + 40 mol% CHOL Laurdan Probe GP−1 3.26 3.35 3.26 3.48 3.62 3.71 3.86 4.20 References [21–24] all performed in liquid environments, but employed different substrates: silicon (Ref. [21]), BSA-coated glass (Refs. [22] and [24]), and APTES-treated mica (Ref. [23]). Regarding the mechanical parameters, Refs. [22] and [23] used the slope of the linear region of the force–deformation curve as the stiffness, Ref. [21] extracted the Young’s modulus based on Shell theory, while Refs. [23] and [24] further quantified the bending modulus (both based on Shell theory) The stiffness of LNPs shown in Table 2 spans an enormous range, extending from the kPa to the GPa level. LNPs with basic ionizable lipids such as MC3, SM-102, and C12-200 are relatively soft, with their stiffness primarily distributed between 63.6 and 118.8 kPa [30]. When formulated with helper lipids into complete mRNA-LNP systems, their stiffness leaps to between 60 and 100 MPa [31]. Furthermore, internal structural alterations, such as metal ion doping, can significantly elevate LNP stiffness to 400–500 MPa [32]. Moreover, introducing a polymer core into LNPs can yield a remarkable stiffness of up to 1.20 GPa [33].Table 2The mechanical properties of LNPs (For mRNA delivery)Mechanical propertiesLNPs (For mRNA delivery)ValueReferencesStiffnessPLGA-LNPs1.20 ± 0.11 GPa[33]PLGA-Water-LNPs0.76 ± 0.07 GPaLNP-mRNA60 MPa[31]L@Mn-mRNA80 MPaMC363.6 ± 17.7 kPa[30]SM-10291.1 ± 32.8 kPaC12-200118.8 ± 61.7 kPaSM-102/DSPC/CHOL/PEG (50/10/38.5/1.5)100 MPa[32]MARF LNP400–500 MPaBending rigidityNot reported to dateMembrane fluidityNot reported to dateReferences [30–33] all reported the Young’s modulus as the mechanical parameter. In terms of substrate, Ref. [30] used a glass slide, while Refs. [31–33] used mica. Refs. [30] and [31] measured in liquid (with Ref. [30] employing a biotin-avidin immobilization strategy), whereas Refs. [32] and [33] measured in air. Refs. [30] and [32] applied the Hertz model, Ref. [33] applied the DMT model, and Ref. [31] did not specify the model employed Table 2The mechanical properties of LNPs (For mRNA delivery)Mechanical propertiesLNPs (For mRNA delivery)ValueReferencesStiffnessPLGA-LNPs1.20 ± 0.11 GPa[33]PLGA-Water-LNPs0.76 ± 0.07 GPaLNP-mRNA60 MPa[31]L@Mn-mRNA80 MPaMC363.6 ± 17.7 kPa[30]SM-10291.1 ± 32.8 kPaC12-200118.8 ± 61.7 kPaSM-102/DSPC/CHOL/PEG (50/10/38.5/1.5)100 MPa[32]MARF LNP400–500 MPaBending rigidityNot reported to dateMembrane fluidityNot reported to dateReferences [30–33] all reported the Young’s modulus as the mechanical parameter. In terms of substrate, Ref. [30] used a glass slide, while Refs. [31–33] used mica. Refs. [30] and [31] measured in liquid (with Ref. [30] employing a biotin-avidin immobilization strategy), whereas Refs. [32] and [33] measured in air. Refs. [30] and [32] applied the Hertz model, Ref. [33] applied the DMT model, and Ref. [31] did not specify the model employed The mechanical properties of LNPs (For mRNA delivery) References [30–33] all reported the Young’s modulus as the mechanical parameter. In terms of substrate, Ref. [30] used a glass slide, while Refs. [31–33] used mica. Refs. [30] and [31] measured in liquid (with Ref. [30] employing a biotin-avidin immobilization strategy), whereas Refs. [32] and [33] measured in air. Refs. [30] and [32] applied the Hertz model, Ref. [33] applied the DMT model, and Ref. [31] did not specify the model employed The stiffness of EVs is highly heterogeneous, reflecting differences in cellular origin and physiological or pathological states [34]. Young’s moduli of EVs from diverse sources range from a few MPa to thousands of MPa [35]. As reported in Table 3, blood- and bone marrow-derived EVs circulating in body fluids are generally softer, typically measuring below 15 MPa [36]. Besides, tumor malignancy is closely correlated with EV mechanics [10, 37]. EVs secreted by highly metastatic or malignant tumor cells typically exhibit reduced stiffness. Additionally, downstream engineering modifications, such as knocking out CD63 in HEK293T cells or lipid hybridization, can effectively fine-tune and remodel the mechanical signatures of EVs [38].Table 3The mechanical properties of EVsMechanical propertiesEVsValueReferencesStiffnessHCV-29 EVs1527 ± 313 MPa[37]T24 EVs95 ± 34 MPaFL3 EVs280 ± 25 MPaF10 EVs30–175 MPa[43]Pan02 EVs25–225 MPaAsPC-1 EVs25–1000 MPa4175 EVs10–200 MPa4T1 EVs10–200 MPaMCF-10A EVs0.5–3.5 × 10−2 N/m[35]MCF-7 EVs0.5–3.5 × 10−2 N/mSK-BR-3 EVs0.5–3.5 × 10−2 N/mMDA-MB-468 EVs0.5–3.5 × 10−2 N/mMDA-MB-231 EVs0.5–6.0 × 10−2 N/mHOS EVs50–200 MPa[44]143B EVs100–350 MPaLM8 EVs50–180 MPaMultiple Myeloma EVs[36]Bone marrow-derived EVs0.5–5 MPaPeripheral blood-derived EVs1–7 MPaLymphoma EVs[36]Bone marrow-derived EVs1–15 MPaPeripheral blood-derived EVs1–8 MPaHealthy blood-derived EVs0.3–3 MPaHEK293T WT EVs0.5–2.2 × 10−2 N/m[38]HEK293T-CD63KO EVs0.1–2.7 × 10−2 N/mHEK293T-Pan-KO EVs0.05–1.7 × 10−2 N/mmilk EVs40 MPa[45]milk EVs@DMPC28 MPamilk EVs@DPPC45 MPamilk EVs@DSPC74 MPaBending rigidityHealthy donor RBC EVs15 kBT[34]Hereditary spherocytosis RBC EVs8 kBTMCF-10A EVs6–31 kBT[35]MCF-7 EVs5–31 kBTSK-BR-3 EVs5–20 kBTMDA-MB-468 EVs4–17 kBTMDA-MB-231 EVs3–18 kBTHEK293T WT EVsHEK293T-CD63KO EVsHEK293T-Pan-KO EVs5–110 kBT5–80 kBT5–75 kBT[38]Membrane fluiditymilk EVsDPH P−1[41]30 °C4.760 °C8.1TMA-DPH r[42]HEK293 EVs0.17A549 EVs0.22Milk EVs0.23References [34–45] (i.e., 34, 35, 36, 37, 38, 43, 44, and 45) differed in substrate, measurement environment, mechanical parameters, and analytical models. In terms of substrate, Refs. [34, 36], and [38] used poly-L-lysine (PLL)-coated glass coverslips. Refs. [35] and [43] used PLL-coated mica. Refs. [37] and [45] used freshly cleaved mica, and Ref. [44] used APTES-modified mica. Refs. [34–38, 43], and [44] all measured in liquid, while only Ref. [45] measured in air. As for the mechanical parameters, Refs. [36, 43, 44], and [45] reported the Young's modulus, Ref. [34] reported the membrane bending modulus, Ref. [37] reported stiffness, Ref. [35] reported both stiffness and bending modulus, and Ref. [38] reported both the effective area modulus and the bending modulus. In terms of analytical models, Refs. [34, 35], and [38] were based on Canham-Helfrich theory, Refs. [36] and [44] applied the Sneddon model, Ref. [37] directly defined stiffness from the slope of the force curve, and Refs. [43] and [45] did not specify the model employed Table 3The mechanical properties of EVsMechanical propertiesEVsValueReferencesStiffnessHCV-29 EVs1527 ± 313 MPa[37]T24 EVs95 ± 34 MPaFL3 EVs280 ± 25 MPaF10 EVs30–175 MPa[43]Pan02 EVs25–225 MPaAsPC-1 EVs25–1000 MPa4175 EVs10–200 MPa4T1 EVs10–200 MPaMCF-10A EVs0.5–3.5 × 10−2 N/m[35]MCF-7 EVs0.5–3.5 × 10−2 N/mSK-BR-3 EVs0.5–3.5 × 10−2 N/mMDA-MB-468 EVs0.5–3.5 × 10−2 N/mMDA-MB-231 EVs0.5–6.0 × 10−2 N/mHOS EVs50–200 MPa[44]143B EVs100–350 MPaLM8 EVs50–180 MPaMultiple Myeloma EVs[36]Bone marrow-derived EVs0.5–5 MPaPeripheral blood-derived EVs1–7 MPaLymphoma EVs[36]Bone marrow-derived EVs1–15 MPaPeripheral blood-derived EVs1–8 MPaHealthy blood-derived EVs0.3–3 MPaHEK293T WT EVs0.5–2.2 × 10−2 N/m[38]HEK293T-CD63KO EVs0.1–2.7 × 10−2 N/mHEK293T-Pan-KO EVs0.05–1.7 × 10−2 N/mmilk EVs40 MPa[45]milk EVs@DMPC28 MPamilk EVs@DPPC45 MPamilk EVs@DSPC74 MPaBending rigidityHealthy donor RBC EVs15 kBT[34]Hereditary spherocytosis RBC EVs8 kBTMCF-10A EVs6–31 kBT[35]MCF-7 EVs5–31 kBTSK-BR-3 EVs5–20 kBTMDA-MB-468 EVs4–17 kBTMDA-MB-231 EVs3–18 kBTHEK293T WT EVsHEK293T-CD63KO EVsHEK293T-Pan-KO EVs5–110 kBT5–80 kBT5–75 kBT[38]Membrane fluiditymilk EVsDPH P−1[41]30 °C4.760 °C8.1TMA-DPH r[42]HEK293 EVs0.17A549 EVs0.22Milk EVs0.23References [34–45] (i.e., 34, 35, 36, 37, 38, 43, 44, and 45) differed in substrate, measurement environment, mechanical parameters, and analytical models. In terms of substrate, Refs. [34, 36], and [38] used poly-L-lysine (PLL)-coated glass coverslips. Refs. [35] and [43] used PLL-coated mica. Refs. [37] and [45] used freshly cleaved mica, and Ref. [44] used APTES-modified mica. Refs. [34–38, 43], and [44] all measured in liquid, while only Ref. [45] measured in air. As for the mechanical parameters, Refs. [36, 43, 44], and [45] reported the Young's modulus, Ref. [34] reported the membrane bending modulus, Ref. [37] reported stiffness, Ref. [35] reported both stiffness and bending modulus, and Ref. [38] reported both the effective area modulus and the bending modulus. In terms of analytical models, Refs. [34, 35], and [38] were based on Canham-Helfrich theory, Refs. [36] and [44] applied the Sneddon model, Ref. [37] directly defined stiffness from the slope of the force curve, and Refs. [43] and [45] did not specify the model employed The mechanical properties of EVs HEK293T WT EVsHEK293T-CD63KO EVs HEK293T-Pan-KO EVs 5–110 kBT 5–80 kBT 5–75 kBT References [34–45] (i.e., 34, 35, 36, 37, 38, 43, 44, and 45) differed in substrate, measurement environment, mechanical parameters, and analytical models. In terms of substrate, Refs. [34, 36], and [38] used poly-L-lysine (PLL)-coated glass coverslips. Refs. [35] and [43] used PLL-coated mica. Refs. [37] and [45] used freshly cleaved mica, and Ref. [44] used APTES-modified mica. Refs. [34–38, 43], and [44] all measured in liquid, while only Ref. [45] measured in air. As for the mechanical parameters, Refs. [36, 43, 44], and [45] reported the Young's modulus, Ref. [34] reported the membrane bending modulus, Ref. [37] reported stiffness, Ref. [35] reported both stiffness and bending modulus, and Ref. [38] reported both the effective area modulus and the bending modulus. In terms of analytical models, Refs. [34, 35], and [38] were based on Canham-Helfrich theory, Refs. [36] and [44] applied the Sneddon model, Ref. [37] directly defined stiffness from the slope of the force curve, and Refs. [43] and [45] did not specify the model employed Membrane bending rigidity Membrane bending rigidity describes the energetic cost required for the local bending of the lipid bilayer, typically denoted by κ in units of kBT or J [39, 40]. This parameter determines the energy barrier for local invagination, budding at the carrier surface, or fusion with the target cell membrane. The bending rigidity of liposomes is closely associated with their phase state and lipid composition (Table 1). Membrane rigidity increases significantly as the phase transitions from the liquid-disordered (Ld) phase to the liquid-ordered (Lo) phase, and ultimately to the solid-ordered (So) phase, with values gradually escalating from 10−19 J to the 10−18 level. Notably, the Lo phase formed by an equimolar mixture of phospholipids and cholesterol typically maintains a rigidity close to 10−19 J. For EVs, this dependency on cellular origin and physiological states extends to the bending rigidity, which is further modulated by membrane protein expression. Pathological conditions can significantly reduce the bending rigidity of EVs. For instance, the rigidity of red blood cell-derived EVs is about 15 kBT in healthy donors, but decreases substantially under specific hematological disorders [34]. Simultaneously, the degree of tumor malignancy is significantly negatively correlated with EV membrane rigidity [35]. Beyond serving as classical biomarkers and maintaining structural integrity, membrane proteins are critical determinants of EV mechanics. Specifically, knocking out the transmembrane protein CD63 significantly reduces their bending rigidity [38]. Membrane fluidity Membrane fluidity was assessed by calculating the generalized polarization (GP or GP−1) of the Laurdan probe, the fluorescence anisotropy (r) of TMA-DPH and the reciprocal polarization (P−1) of DPH. TMA-DPH anisotropy increases from 0.195 to 0.265 upon adding 50 mol% cholesterol to unsaturated EPC, reflecting reduced fluidity. In contrast, increasing cholesterol from 10 to 50 mol% in DSPC lowers the anisotropy from 0.288 to 0.187, indicating a fluidizing effect [25]. Saturated phospholipids such as DSPC, DPPC, and DMPC form compact membranes with high GP values ranging from 0.47 to 0.50, whereas unsaturated lipids like DOPC, DOPE, and POPC lower these values to 0.40–0.43, thereby enhancing fluidity [26]. Furthermore, heating DSPC, cholesterol, and DSPE-PEG2000 systems from 37 to 45 °C increases GP−1 values from 3.26–3.48 to 3.62–4.20, markedly boosting fluidity. Under identical conditions, glycerol isotonic solutions consistently exert a stronger fluidizing effect than glucose. Finally, increasing the cholesterol proportion from 20 to 40 mol% at 45 °C further augments the membrane fluidity of this system [27]. EV membrane fluidity is governed by both temperature and cellular origin (Table 3). Based on the P−1 value of the DPH probe, the fluidity of milk exosomes exhibited only a slow and gradual increase, from 4.7 at 30 °C to 8.1 at 60 °C [41]. EVs from different sources exhibit minimal differences in fluidity, with their anisotropy values primarily ranging from 0.22 to 0.24. In comparison with model membranes, these values indicate the coexistence of Lo and Ld phases within their membranes [42]. Currently, research attention across the three key mechanical parameters is highly uneven. In particular, studies on the membrane bending rigidity and fluidity of mRNA-LNPs remain exceedingly rare. However, the significant heterogeneity exhibited in the mechanical properties of LNCs fundamentally stems from their underlying molecular architecture. To truly achieve a rational design of the mechanical properties of LNCs, it is imperative to look beyond superficial physical parameters and deeply investigate the material basis that dictates these characteristics. Structural determinants of mechanical properties Until now, we understand that the mechanical properties of LNCs are mainly dictated by three structural determinants, including the organization of the lipid, cholesterol abundance, and internal architecture (Fig. 1).Fig. 1Compositional and structural origin of lipid nanocarrier (LNC) mechanical properties. a, the organization of the lipid. b, mechanical buffering of cholesterol. c, influence of internal structure. Schematic conclusions were drawn from Refs. [46–49] Created in BioRender. Chiappini, C. (2026) https://BioRender.com/eu3zhaa Compositional and structural origin of lipid nanocarrier (LNC) mechanical properties. a, the organization of the lipid. b, mechanical buffering of cholesterol. c, influence of internal structure. Schematic conclusions were drawn from Refs. [46–49] Created in BioRender. Chiappini, C. (2026) https://BioRender.com/eu3zhaa The organization of the lipid Governed by membrane biophysics, the microscopic molecular structure of LNCs strictly regulates their thermodynamic state, the phase transition temperature Tm, ultimately determining their mechanical properties. Tm is primarily dictated by the phospholipid acyl chains. Driven by higher transition temperatures, long-chain saturated phospholipids form an So phase at physiological conditions, whereas short-chain unsaturated ones favor a fluid Ld phase [46, 47]. With the Lo phase acting as an intermediate state, the membrane’s fluidity follows an increasing order of So < Lo < Ld, whereas its bending modulus decreases as So > Lo > Ld. Liposomes possess well-defined compositions and highly controllable mechanical properties. Varying only the helper phospholipid reveals that unsaturated lipids yield significantly higher membrane fluidity than saturated ones [26]. The membrane’s bending rigidity increases monotonically with acyl chain length. Conversely, introducing cis-double bonds disrupts ordered chain packing, progressively reducing bilayer bending rigidity as more double bonds are added [39]. Furthermore, introducing charged lipids decreases membrane rigidity by 30–60% [22]. Unlike liposomes, EVs are enriched in saturated sphingomyelin and cholesterol. Their highly variable lipid composition across cell sources and even subpopulations directly leads to mechanical disparities. Small EVs (sEVs) are stiffer due to small size and dense molecular packing driven by flotillin and tetraspanins, whereas larger EVs (lEVs) are softer, predominantly enriching membrane-associated proteins like annexins [43]. Membrane protein networks critically regulate EV elasticity. Notably, knocking out CD63 alters membrane cholesterol, fundamentally shifting the mechanical properties of EVs [38]. Mechanical buffering of cholesterol Embedded between phospholipid tails, cholesterol acts as a mechanical buffer, exerting a bidirectional regulatory effect on both membrane rigidity and fluidity [50, 51]. In the Ld phase, cholesterol restricts acyl chain mobility, condensing the membrane to increase stiffness. In contrast, in the So phase, it disrupts ordered lipid packing, thereby decreases rigidity and increases fluidity [23, 24]. At physiological temperatures, cholesterol tends to induce a Lo phase, where lipid chains remain ordered while retaining lateral diffusion, thereby maintaining membrane fluidity and mechanical stiffness at a moderate, stable level across a broad temperature range [25, 47, 51, 52]. Influence of internal structure Internal structure fundamentally influenced the mechanical distinctions among liposomes, LNPs and EVs [49]. Unlike traditional liposomes, whose mechanics rely almost entirely on the outer lipid bilayer due to a simple aqueous core, EVs possess a complex internal network of soluble proteins, cytoskeletal fragments, and nucleic acids. Metastatic tumor cells expel elastic fiber-associated proteins into EVs to maintain cellular softness for invasion, resulting in metastatic EVs exhibiting a significantly higher Young’s modulus than non-metastatic ones [44]. For LNPs, the ionizable lipids and nucleic acids form a dense core that significantly increases overall rigidity. Payload type and conformation crucially dictate LNPs’ mechanics. Specifically, pDNA-LNPs are stiffer than mRNA-LNPs, exhibiting rupture forces of 500 pN and 200 pN, respectively [49]. Moreover, metal ions such as Mn2+ can induce mRNA to fold into compact structures, acting as an internal rigid scaffold that drastically enhances the overall elastic modulus of the LNP [31, 32]. The stiffness is also modulated by the lipid formulation, particularly the structural variations of sterol analogs [30]. In core–shell NPs, while a PLGA core provides basal rigidity, an interfacial water layer can act as a fluid buffer to significantly reduce the overall hardness of the NPs [33]. Among LNCs, liposomes serve as the most ideal physical entities for validating and applying classical membrane mechanics theories. In contrast, the outer layer of LNPs is a disordered mixed layer rather than a uniform bilayer, limiting the applicability of conventional membrane mechanics theories and characterization techniques. Although the theoretical mapping among composition, structure, and mechanics has been well established, accurately capturing these weak physical signals at the nanoscale relies heavily on advanced characterization techniques. Characterization techniques Atomic force microscopy The quantification of mechanical properties of LNCs is a prerequisite for their controllable design. Atomic Force Microscopy (AFM) is a pivotal tool for evaluating the morphology and mechanics of LNCs. By analyzing force-distance (FD) curves generated during vertical indentation, AFM quantitatively extracts nanomechanical parameters such as Young’s modulus and deformability [34, 53–57]. Its label-free, in situ operation provides single-particle resolution, simultaneously capturing morphological and mechanical traits to reveal sample heterogeneity [58, 59]. Despite its high spatial and mechanical resolution, the reliability and reproducibility of AFM force measurements on LNCs remain inherently challenging, owing to multiple technical variables spanning the entire experimental workflow from sample preparation through data acquisition to analysis [60]. This has hindered the development of standardized, cross-laboratory comparable datasets. The paradox between the ubiquitous use of AFM for LNC mechanical profiling and the persistent difficulties in standardization highlights the critical need to explore the underlying factors affecting measurement outcomes. Achieving reliable measurements begins at the sample preparation stage, where PLL-coated mica substrates under physiological conditions have become widely adopted to preserve the native three-dimensional (3D) structural and mechanical integrity of LNCs [21]. However, this electrostatic adsorption process itself inevitably induces a degree of particle deformation at the substrate interface. It is reported that this deformation can be assessed by quantifying the contact angle of adsorbed particles following AFM imaging [61]. Beyond substrate effects, probe-related variables introduce further complexity. Tip geometry and radius of curvature directly modulate force measurements [23]. A larger tip apex enlarges the contact area, requiring greater energy for membrane deformation and consequently overestimating interaction forces, whereas a sharper tip concentrates stress over a reduced contact area and risks membrane rupture under high loads. When force measurements and morphological imaging are conducted simultaneously, tip selection therefore demands a careful balance between mechanical fidelity and imaging resolution. The cantilever spring constant exerts a pronounced influence on measurement outcomes. A stiffer cantilever imparts greater actual force per unit piezo displacement, inducing larger deformation and steeper force–distance curves that systematically elevate the apparent elastic modulus. Tips with excessively high spring constants potentially drive membrane puncture and yield substantial artifactual overestimations [62]. The choice of contact mechanics model represents the most consequential analytical factor. The widely used Hertz contact model treats the sample as a homogeneous, isotropic, semi-infinite elastic solid. The physical reality of nanoscale lipid particles fails to meet these assumptions, leading to significant underestimation of membrane bending stiffness [23, 34]. The Shell model is structurally more representative of LNCs as hollow, membrane-enclosed particles, making it a better-suited framework for their mechanical characterization [63]. Recently, to address the inter-laboratory variability in reported values, an optimal standardized protocol was recommended using cantilever spring constants of 0.02–0.08 N/m, tip radii below 15 nm, an indentation depth of 5–10% of the vesicle diameter, substrate functionalization with PLL or APTES [56]. Where the goal is to obtain absolute modulus values comparable across laboratories and measurement platforms, a two-step approach, combining low-force morphological imaging with high-force indentation, enables particle-level extraction of radius of curvature, compressive stiffness, and membrane thickness, from which the apparent Young’s modulus can be resolved. The Hertz contact model, by contrast, remains a practical and widely adopted alternative in soft vesicle studies due to its simpler workflow. Under matched experimental conditions, systematic errors cancel across groups, preserving the statistical validity of inter-group comparisons. It is therefore well suited for studies where relative mechanical differences, rather than absolute values, are the primary outcome of interest [63]. Membrane fluidity probes Fluorescence anisotropy and generalized polarization based on DPH, TMA-DPH, and Laurdan probes enable a depth-dependent characterization of membrane fluidity, spanning from the hydrophilic surface to the hydrophobic core [27, 64]. Rigid membranes yield high anisotropy and GP values. As a non-destructive, high-throughput method, it effectively monitors fluidity changes induced by cholesterol content, lipid tail saturation, or temperature. These techniques establish the precise mechanical mapping of LNCs. However, acquiring static mechanical parameters is merely the starting point. Their ultimate scientific value lies in deciphering how these mechanical properties dictate in vivo fate of the carriers, acting as crucial regulators at the bio–nano interaction (Fig. 2).Fig. 2Representative examples of mechanical regulation of LNC behavior at the bio–nano interface. a, Effect of mechanical properties on protein corona formation. b, Effect of mechanical properties on membrane wrapping. c, Effect of mechanical properties on internalization. Schematic conclusions are drawn from Refs. [65–67]. Created in BioRender. Chiappini, C. (2026) https://BioRender.com/p9dgk1s Representative examples of mechanical regulation of LNC behavior at the bio–nano interface. a, Effect of mechanical properties on protein corona formation. b, Effect of mechanical properties on membrane wrapping. c, Effect of mechanical properties on internalization. Schematic conclusions are drawn from Refs. [65–67]. Created in BioRender. Chiappini, C. (2026) https://BioRender.com/p9dgk1s Mechanical properties of LNCs regulating bio–nano interaction Effect of mechanical properties on protein corona formation In biological fluids, LNCs rapidly acquire a protein corona comprising a hard corona of tightly and stably bound proteins and a soft corona of loosely associated proteins in dynamic exchange with the surrounding medium. NPs mechanics have been reported to influence protein adsorption patterns through deformability and stress responses at the interface, yet studies directly correlating quantitative mechanical data of LNCs with differences in protein corona formation remain limited [68]. It was reported that a liposome-sodium alginate hydrogel NPs (LSHNPs) with enhanced elasticity was developed, showing that the more elastic LSHNPs preferentially enriched apolipoprotein E (ApoE) and albumin, whereas the less elastic LSPNPs selectively adsorbed fibrinogen α chains [69]. Among the three groups of LNPs, classified as Soft, Semi-rigid, and Hard based on their respective Young’s moduli, no statistically significant differences were detected in the total protein content of the adsorbed colitis-specific intestinal protein corona (C-IPC). However, high-rigidity NPs were capable of selectively adsorbing a greater proportion of macrophage-targeting proteins, with the proportion of S100A8 protein being most notably upregulated [70]. However, direct evidence linking the mechanical properties of EVs to their protein corona formation remains scarce in the current literature. Given that LNCs share similar lipid-based structures, it has been proposed that the formation of the protein corona on the LNCs proceeds through a three-stage screening mechanism involving headgroup charge, membrane fluidity, and bending rigidity [71]. In biological fluids, initial protein adsorption onto liposomes is primarily driven by electrostatic interactions with lipid headgroups [47]. Lipid headgroups selectively bind specific proteins. For example, cationic lipids preferentially adsorb negatively charged plasma proteins, shifting the carrier’s surface potential to between − 1 and − 30 mV [72]. PEGylation generally reduces corona thickness without fundamentally altering its protein composition [73]. After initial electrostatic binding, membrane fluidity is associated with the protein penetration into the hydrophobic core [65]. Densely packed membranes, composed of saturated lipids, restrict protein binding primarily to lipid packing defects. Consequently, proteins remain confined to the lipid-water interface, forming a highly dynamic soft corona. In contrast, Ld membranes feature abundant packing defects that allow deep insertion of hydrophobic domains, yielding a hard corona with higher binding affinity and protein capacity (Fig. 2a). Membrane bending rigidity has been proposed to act as the third mechanical determinant that may influence the retention of conformation-sensitive proteins and the carrier’s structural integrity. Rigid membranes may disfavor the adsorption of proteins whose binding requires substantial membrane deformation, yet may preferentially enrich certain conformation-sensitive proteins such as S100A8, as reported in one study [70]. Flexible membranes can deform more readily, which has been suggested to enrich the compositional diversity of the protein corona. This diversity may be greatest particularly during microphase separation near the lipid phase Tm [65]. While it is increasingly recognized that LNC mechanics shape corona composition and architecture, how the corona in turn modifies LNC mechanical properties remains to be explored. Direct experimental evidence remains sparse, with only a handful of studies showing that protein corona influenced LNC mechanics in a carrier and preparation method-dependent way [38, 74]. Effect of mechanical properties on membrane wrapping Because LNC membrane wrapping occurs at nanoscale and millisecond spatiotemporal scales, direct high-resolution observation is extremely challenging, and direct quantitative mechanical data are likewise difficult to obtain. Studies from related model systems may provide useful insights. Based on theoretical calculations using the Helfrich model, the energy changes during membrane wrapping were found to differ fundamentally between soft NPs and rigid NPs [75]. While the following studies were conducted on non-LNC systems, they provide mechanistic principles that are likely applicable to LNCs. Using a combination of transmission electron microscopy (TEM), flow cytometry, and theoretical calculations, it was found that the flattened soft silica nanocapsules (SNCs) occupied more folate receptors. The excessive deformation of soft SNCs increased the energy barrier required for complete membrane wrapping, consequently lowering their internalization rate (Fig. 2b) [66]. Simulations revealed that the LNC-to-membrane bending modulus ratio (κ1/κ2) governs these interfacial dynamics. Hard NPs (κ1/κ2 ≫ 1) induce active membrane wrapping with a minimal energy cost (ΔE). Soft NPs (κ1/κ2 ≪ 1) tended to deform, and simulations indicated that this can raise the thermodynamic barrier for complete wrapping by approximately 30%, which may delay internalization [33, 76, 77]. In macrophage phagocytosis of microparticles, it was found that stiff cargo is rapidly engulfed, while soft cargo induces central F-actin protrusions and phagocytic stalling [78]. Taken together, although these findings derive from non-LNC model systems, they suggest that a low κ1/κ2 ratio may increase energy barriers and reduce receptor-mediated actin recruitment, which could in turn contribute to the lower cellular uptake of soft LNCs and has been associated with prolonged in vivo circulation in some studies. Effect of mechanical properties on internalization Across the reported systems, LNC cellular uptake has been influenced by the interplay between NP mechanics and physiological barriers. When using natural-lipid-based N-LNPs as a model system, surface fluidity and stiffness were identified as co-determinants of cellular uptake vitality: soft N-LNPs with high surface fluidity (N-LNPs_I) demonstrated preferentially higher uptake by HeLa cells, whereas stiffened N-LNPs (hardened by dsDNA encapsulation) shifted toward clathrin-mediated endocytosis as the dominant internalization pathway, as evidenced by enhanced sensitivity to chlorpromazine (CPZ) inhibition [79]. Soft EVs (low stiffness) were internalized in significantly greater quantities by both tumor and non-tumor cells compared to their stiff counterparts; however, this trend is not absolute. Within the high stiffness regime (~ GPa, e.g., PLGA-based nanoparticles), cellular uptake increased with stiffness, suggesting a non-monotonic, biphasic relationship between particle stiffness and internalization efficiency [67]. LNPs incorporating unsaturated phospholipids exhibited higher membrane fluidity, as quantified by the Laurdan method, and demonstrated enhanced cellular uptake across multiple cell types compared to their saturated counterparts [26]. Beyond uptake efficiency, the mechanics of LNCs have been associated with shifts in the primary cellular endocytosis pathways, although the reported associations are system-dependent. In the mExos-liposomes hybrid vesicle model, soft vesicles (mExos@DMPC) were internalized by Caco-2 cells predominantly via clathrin-mediated endocytosis and micropinocytosis, whereas hard vesicles (mExos@DSPC) relied primarily on caveolin-mediated endocytosis, indicating that vesicle stiffness influences the endocytic pathway engaged during cellular internalization [45]. The stiffness of EVs influences their internalization pathway in a biphasic manner: soft EVs in the low stiffness regime (~ MPa) preferentially enter cells via membrane fusion, whereas stiff EVs in the high stiffness regime (~ GPa) shift toward endocytosis as the dominant internalization pathway [67]. These mechanics-dependent uptake behaviors are thought to arise from energetic and structural interactions between the LNCs, cell membranes, and the surrounding microenvironment. The structural stability of these rigid carriers may also enhance the recognition and binding affinity between surface proteins and cellular receptors [45]. Collectively, the association between carrier stiffness and endocytic pathway depends on both the carrier type and the stiffness regime in question. Engineering strategies for mechanical modulation This section reviews engineering strategies through which the mechanical properties of liposomes, LNPs, and EVs can be rationally modulated, and examines how the resulting mechanical changes alter their behavior at the bio–nano interface. The three platforms arise from different structural principles and therefore offer different mechanical design strategies, so they are analyzed separately below (Fig. 3). For each strategy we first identify how it changes the mechanical properties, distinguishing directly measured mechanical properties from proxy or inferred measures, and how the mechanical properties affect the biological functional outcomes. Because mechanics is still an emerging design axis for these carriers, we also flag where a mechanical link is asserted but not yet directly measured.Fig. 3Engineering strategies for mechanical modulation of liposomes, LNPs, and EVs. Created in BioRender. Chiappini, C. (2026) https://BioRender.com/6iv27mz Engineering strategies for mechanical modulation of liposomes, LNPs, and EVs. Created in BioRender. Chiappini, C. (2026) https://BioRender.com/6iv27mz Mechanical modulation of liposomes Here, four strategies to modulate liposome elasticity are outlined: compositional engineering and thermodynamic phase regulation. Compositional engineering: phospholipid As mentioned previously, the acyl chain length and degree of unsaturation of phospholipids are important structural determinants of mechanics, primarily by modulating the Tm. Lipids with higher-Tm produce stiffer membranes, while lower-Tm lipids produce softer ones [8]. Dai et al. systematically quantified this across six SUV formulations spanning C12:0–C22:1 phosphatidylcholines [80]. The elastic modulus E increased monotonically from 5.8 MPa (DLPC, C12:0) to 42.8 MPa (DSPC, C18:0) for saturated species, while introduction of a single cis double bond (DOPE, C18:1) reduced E to levels comparable to the shortest saturated chain despite the longer backbone. Functionally, the stiffest formulation (DSPC, E = 42.8 MPa) showed the highest cellular internalisation in vitro via clathrin-mediated endocytosis. However, this performed poorly in vivo because stiff vesicles could not deform within the fibrous ECM and failed to penetrate tumor stroma. Intermediate-rigidity DPPC liposomes (E≈20 MPa) achieved 11.8-fold higher ECM diffusivity than the softest group, penetrated deepest into multicellular spheroids, and retained 63% of the injected dose at the tumor site at 24 h compared to 19–42% for softer or stiffer counterparts, achieving the optimized trade-off. STED microscopy directly visualized the mechanism: DPPC liposomes underwent controlled spherical-to-ellipsoidal deformation within ECM fibers (aspect ratio of 2.5), a geometry that minimizes adhesive drag while maintaining structural integrity for subsequent endocytosis. These findings establish that phospholipid identity does not simply scale delivery performance with stiffness, but defines an optimal mechanical window determined by the competing demands of stromal penetration and cellular uptake. Compositional engineering: cholesterol Cholesterol inserts between phospholipid acyl chains and increases membrane bending rigidity κ. Chakraborty et al. used neutron spin-echo (NSE) spectroscopy, solid-state 2H NMR, and molecular dynamics simulations to show that cholesterol increases κ of DOPC membranes by increasing the bilayer’s packing density [81]. Liu et al. extended this framework by substituting native cholesterol with plant-derived analogues that differ in their C3-position functional groups, including free sterol (FS), sterol ester (SE), and steryl glycoside (SG), while maintaining the same steroid backbone [82]. Substituents with greater hydrogen-bonding capacity at C3 increased lateral chain packing compared to native cholesterol in DPPC nanoliposomes, raising E as measured by AFM nanoindentation (in the order of P-SG > P-SE > P-FS > P–CH). This mechanical enhancement translated directly into higher oral bioavailability of encapsulated oligomeric proanthocyanidin in vivo, with the stiffest formulation (P-SG-OPC) exhibiting the greatest melanin-inhibiting efficacy in zebrafish, demonstrating cholesterol structure can be used to tune both mechanical properties and in vivo bioavailability within a fixed phospholipid framework [82]. Compositional engineering: surface modification Surface modification is routinely employed in nanoformulation design, yet its mechanical consequences are underappreciated relative to those of bulk lipid composition. PEGylation, primarily adopted for its stealth properties, also produces measurable and functionally relevant mechanical changes that scale with PEG chain length and grafting density. Mahendra et al. showed that the bending modulus increases with PEG mole fraction and that this stiffening is more pronounced for longer chains. The area compressibility modulus remained unchanged across all conditions, indicating that PEGylation selectively stiffens the membrane against bending without altering in-plane stretching resistance. Functionally, rising PEG content increased water permeability and decreased lysis tension. Moreover, as the PEG content increases, it can also predict that the leakage amount of the loaded substance in liposomes with highly PEG-modified surface will increase [83]. Jahnke et al. further examined the mechanical effects of chondroitin sulfate (CS) functionalization [84]. Using micropipette aspiration, they demonstrated that cholesterol-functionalized chondroitin sulfate (Chol-CS) attachment reduced the stretching modulus KA by ~ 15% and decreased bending rigidity from 17 to 13 kBT, while these effects were absent in cholesterol-tethered PEG controls. The softening mechanism contrasts with the previously reported PEG-mediated stiffening [83]: the high negative charge density of CS drives electrostatic repulsion between neighboring Chol-CS chains, increasing their intermolecular separation and reducing the membrane’s resistance to lateral stretching, a mechanism unavailable to the uncharged PEG chain. The downstream delivery consequences of this polysaccharide-induced bilayer softening remain to be explored in vivo. It is worth noting that negatively charged biopolymers commonly employed as active-targeting ligands, such as hyaluronic acid and heparin, share this polyanionic architecture, raising the possibility that ligand-mediated bilayer softening may represent an underappreciated variable in active-targeting nanoparticle formulations that warrants systematic investigation. Thermodynamic phase state Below Tm, phospholipid acyl chains predominantly adopt all-trans conformations and pack into the solid-ordered So phase. Above Tm, trans-gauche rotational isomerization introduces chain disorder and kinks, and the membrane enters the liquid-crystalline phase (Lα). Near Tm, an anomalous decrease in bending rigidity has been observed, attributed to increased monolayer compressibility [85]. Yu et al. prepared liposomes with Tm values spanning 16–55 °C by varying lipid composition (spanning DOPE, DMPC, DPPC, HSPC, and DSPC-based formulations) and measured diffusivity in PEO hydrogel and rat intestinal mucus across multiple temperatures (4–55 °C) [86]. Liposomes with Tm near ambient temperature showed the fastest transport. Gel-phase liposomes with Tm higher than ambient temperature maintained their spherical shape and became trapped at network junctions due to high polymer-lipid attraction density; while fully fluid liposomes (Tm << ambient temperature) conformed to and became trapped within the hydrogel network. Functionally, near-Tm liposomes achieved a maximal blood glucose reduction of 50% persisting for over 8 h and a relative insulin bioavailability of 13.65% following oral administration in diabetic rats, significantly surpassing both fully fluid and gel-phase counterparts, supporting ambient-temperature-matched Tm as a rational and actionable formulation principle for navigating biological hydrogel barriers. Mechanical modulation of LNPs Several FDA-approved nucleic acid therapeutics, including Onpattro, Comirnaty, and Spikevax, share a common LNP architecture where four components (an ionizable lipid, helper phospholipid, cholesterol, and PEG-lipid) self-assemble around the payload [87]. The ionizable lipid electrostatically condenses the nucleic acid cargo into a solid interior matrix whose structural organization contributes to both particle mechanics and delivery competence. LNP mechanical properties involve both bulk stiffness, dictated by lipid composition, and the internal nanostructure phase state [30]. Specifically, the liquid-crystalline organization of the core governs pH-triggered mesophase transitions and endosomal escape [88]. These two parameters are complementary: Bulk stiffness reflects the macroscopic deformability of the intact NP, while internal phase state describes the structural reorganization required for intracellular cargo release. The following subsections address how each component and preparation parameter modulates one or both levels. Ionizable lipid structure The ionizable lipid is the largest contributor to LNP core mass, and its tail architecture determines the packing density and dynamic mobility of the lipid interior. Hashiba et al. synthesized a library of 32 ionizable lipids and showed that tail branching increases core microviscosity [89]. By optimizing tail branching, the optimal formulation achieved 54% genome editing and 77% protein reduction at 2.5 mg kg−1 in a CRISPR–Cas9 model [89]. Core microviscosity is a nanoscale rheological parameter reflecting resistance to lipid chain displacement, and higher values may imply greater whole-particle stiffness, although AFM-based Young’s modulus measurements were not reported in this study. Mrksich et al. varied tail length within the C12-200 homologue series and found that shorter tails enhanced the delivery of larger mRNA cargoes, with C10-200 increasing liver luciferase expression over tenfold relative to C12-200 [90]. Across these and related studies, no investigation has directly measured the effect of ionizable lipid structure on AFM-derived Young’s modulus, representing a clear knowledge gap. Establishing this link would connect the extensive ionizable lipid structure–activity literature to the emerging mechanical framework for LNP design. Helper phospholipid Helper phospholipids constitute the outer structural layer of LNPs and govern the balance between colloidal stability during circulation and membrane fusogenicity at the endosome. DSPC, a saturated phosphatidylcholine with a melting temperature of 54 °C, adopts a cylindrical geometry (CPP≈1) that promotes lamellar phase formation, conferring a rigid, ordered outer shell resistant to premature disassembly [1]. DOPE, a phosphatidylethanolamine with two unsaturated oleic acid tails, undergoes a hydration- and measurement-dependent lamellar-to-HIItransition well below physiological temperature and adopts a conical molecular shape (CPP > 1); at physiological temperature it readily forms inverted hexagonal (HII) phases that destabilise endosomal membranes and facilitate cargo escape [1]. These contrasting phase behaviors carry direct consequences for LNP structural dynamics across the delivery pathway. Li et al. applied small-angle neutron scattering (SANS) with isotopic contrast variation to track the acidification-induced structural evolution (AISE) of pDNA-containing LNPs formulated with five phosphatidylcholine helper lipids spanning a range of saturation states (SOPC, POPC, DMPC, DPPC, and DSPC) [91]. At pH 7.4, all helper lipids tested preferentially localized to the LNP outer shell, supporting a bilayer outer membrane. Upon acidification to pH 4.0 to mimic endosomal maturation, LNPs universally underwent AISE: volume expansion accompanied by water ingress and redistribution of lipid components. LNPs containing unsaturated helper lipids (SOPC, POPC) exhibited greater AISE than their saturated counterparts (DMPC, DPPC, DSPC): unsaturated helper lipids were more readily redistributed from the outer shell into the LNP interior, increasing outer-shell hydration from ~ 25 to ~ 48% and promoting non-bilayer structural defects that facilitate endosomal membrane fusion. Functionally, unsaturated helper lipids achieved up to 15-fold higher in vitro luciferase expression in HeLa cells than saturated analogues. Zhang et al. further demonstrated that DOPE-LNPs formed stronger, more rigidly coupled surface interactions with ApoE than DSPC-LNPs as measured by quartz crystal microbalance with dissipation (QCM-D), correlating with preferential liver accumulation and higher hepatic transfection [92]. Direct AFM nanoindentation measurements of whole-particle Young’s modulus as a function of helper lipid identity have not been reported. The mechanical dimensions evidenced here, including gel/fluid outer-shell phase rigidity, pH-responsive structural reorganization amplitude (AISE), and interfacial adsorption coupling (QCM-D), are therefore complementary proxies for helper-lipid-dependent mechanics, each operating on a distinct length scale and at a different intracellular stage. Cholesterol Cholesterol modulates membrane rigidity and LNP structural integrity and is currently the best-characterized tool for tuning whole-particle stiffness, with downstream consequences for both NP mechanics and intracellular delivery [30, 93]. Patel et al. systematically substituted cholesterol with naturally occurring C-24 alkyl phytosterols in MC3-based LNPs; the resulting LNPs adopted highly faceted, polyhedral morphologies by cryo-TEM and achieved up to 48-fold improvements in mRNA transfection across multiple cell lines and ionizable lipid identities [93]. The authors attributed this to NPC1/NPC2-mediated endosomal retention and altered surface lipid composition, although the mechanical properties of the phytosterol-substituted LNPs were not measured. Building on this, Safford et al. quantified the mechanical consequences of phytosterol substitution in a series of C12-494 LNPs by AFM [30]. Young’s modulus increased from 71.0 ± 26.2 kPa with native cholesterol to 201.7 ± 71.4 kPa with β-sitosterol and 411.4 ± 145.7 kPa with stigmastanol. In vitro screening in placental trophoblasts showed that the intermediate stiffness (β-sitosterol of 202 kPa) significantly improved uptake and mRNA delivery, whereas the stiffest formulation (stigmastanol of 411 kPa) did not sustain this advantage. Following intravenous injection in pregnant mice, β-sitosterol LNPs enhanced placental mRNA expression and reduced liver mRNA delivery relative to cholesterol controls [30]. This intermediate-rigidity optimum parallels the pattern observed for liposomes in tumor ECM [80], although whether this represents a generalizable design principle across LNP compositions and target organs remains to be determined. Cargo identity The encapsulated nucleic acid cargo represents an additional mechanical variable beyond the lipid composition discussed above. De Chateauneuf-Randon et al. directly tested this by formulating pDNA-LNPs and mRNA-LNPs at identical lipid compositions and a constant N/P ratio of 6. AFM indentation revealed that pDNA-LNPs displayed a rupture-force distribution centered around 500 pN and were irreversibly damaged after indentation, whereas mRNA-LNPs required lower forces of 150–200 pN and frequently showed partial recovery after tip retraction. The authors attributed these differences to the greater intrinsic rigidity of double-stranded pDNA relative to single-stranded mRNA and the resulting differences in the polyanion–lipid complex organization. These findings support the view that cargo identity modulates the internal mechanical state of LNPs, with potential consequences for how they deform and reorganize during interactions at the cellular interface [49]. Mechanical engineering of EVs Cell source as an intrinsic mechanical determinant As outlined in Sect. 2, the cell-source-dependent mechanical properties of EVs dictate their bio–nano interactions, originating from specific biogenesis processes such as lipid sorting, vesicle budding, and cargo loading. EVs derived from cancer cells provide a well-characterized example of the relationship between parent cell identity, vesicle molecular composition, mechanical properties, and biological function [94, 95]. Whitehead et al. found that malignant bladder cancer EVs are markedly softer and less adhesive than non-malignant counterparts. Functionally, these softer malignant EVs dose-dependently disrupted endothelial monolayers and induced stronger complement activation, effects that were absent or muted in non-malignant EVs [37]. Ye et al. demonstrated that the apparent stiffness of malignant cell-derived sEVs results from the combined contribution of the membrane bending modulus and the osmotic pressure [35]. Breast cancer sEVs exhibited malignancy- and size-dependent mechanical alterations, where overall stiffness increased with malignancy. This stiffening directly contrasts with the softening observed in bladder cancer exosomes, reflecting distinct, cancer-type-specific mechanical adaptations [37]. Even within a single tumor cell line, secreted EV subpopulations exhibit distinct mechanical profiles, with sEVs being notably stiffer than larger ones. Crucially, these mechanically distinct subpopulations display different in vivo organ biodistribution, indicating that mechanical identity contributes to subtype-specific tissue targeting [43]. This cell-source-dependent mechanical heterogeneity extends beyond cancer to therapeutically relevant non-malignant cell types. Cavallaro et al. combined single-vesicle AFM with fluorescence microscopy to study sEVs from cord blood mesenchymal stromal cells (cbMSCs), HEK293 and monocytic leukemia cells. Within each cell line, sEVs resolved into subpopulations with distinct mechanical properties and surface marker profiles. Interestingly, one subpopulation with average diameters below 100 nm and high co-expression of all three tetraspanins (CD9, CD63, CD81) consistently exhibited low Young’s moduli across all three cell lines, regardless of cellular origin, suggesting a potentially conserved biogenesis pathway linking surface marker identity to the mechanical properties of the vesicles [96]. Collectively, these studies demonstrate that the mechanical properties of EVs are cell-source-dependent and subpopulation-specific, with no universal stiffness trend even within cancer cell lines [35, 37]. Within a single secretome, biogenesis sorting generates mechanically distinct vesicle subpopulations with different biodistribution profiles. Cell source therefore determines both the molecular cargo and the mechanical properties of the EVs, which together govern their downstream interactions with biological barriers and recipient cells. Post-biogenesis engineering modifications As demonstrated in Sects. 3.1 and 3.2 for liposomes and LNPs, direct membrane modifications can also be applied to EVs. By incorporating cholesterol into milk-derived EVs, Zhuo et al. significantly softened the vesicles to increase their deformability and cell-membrane contact area. Functionally, above 23 wt% cholesterol, the dominant uptake route shifted from endocytosis to direct membrane fusion, enabling cytosolic delivery that bypassed lysosomal degradation. The 30 wt% formulation achieved 84% PLK1 siRNA-mediated knockdown and produced tumor regression in an orthotopic colorectal cancer model following oral administration, outperforming both native EVs and LNP comparators. Interestingly, this endocytosis to fusion transition was cancer-cell-selective, with normal cells internalized the same cholesterol-modified vesicles exclusively via endocytosis regardless of cholesterol content, suggesting that recipient cell membrane properties also modulate fusion-mediated entry [15]. Other surface-engineering strategies have also been applied to modify EV surfaces, although without directly targeting membrane mechanics. Kooijmans et al. decorated isolated EVs with EGFR-targeting nanobodies conjugated to DMPE-PEG–phospholipid micelles via temperature-dependent post-insertion [48]. The introduction of nanobody-PEG-lipid constructs onto EV membranes extended plasma circulation time from 10 min for unmodified controls to beyond 60 min. While PEGylation alone suppressed non-specific binding to non-target cells (Neuro2A), decoration with EGFR-specific EGa1 nanobodies selectively enhanced binding to EGFR-overexpressing A431 tumor cells in vitro, demonstrating that stealth properties and active targeting can be simultaneously achieved through this strategy [48]. Lathwal et al. further engineered the EV surface with synthetic polymers (pOEOMA, pCBMA, and pMSEA) via atom transfer radical polymerization (ATRP) through cholesterol-anchored DNA tethers [97]. These EV-polymer hybrids exhibited a fourfold extension of blood-circulation half-life compared with native EVs, while preserving the intrinsic angiogenic activity of MSC-derived EVs and the osteogenic activity of BMP2-loaded EVs in vitro [97]. However, neither study characterized how these surface modifications altered EV membrane stiffness, leaving it unresolved whether mechanical changes at the vesicle surface contributed to, or were independent of the observed shifts in biodistribution and cell-entry behaviors. In addition to surface modifications, EV-liposome hybridization enables systematic tuning of membrane mechanics with matched physicochemical properties. Xiao et al. fused milk EVs with phospholipids of differing acyl-chain lengths, including DMPC (C14), DPPC (C16), or DSPC (C18), via dried film hydration and co-extrusion, achieving a tunable mechanical gradient while controlling for size, charge, and surface hydrophilicity [45]. Consistent with previous findings, the softer hybrids penetrated intestinal mucus most efficiently (3.2-fold greater mean-square displacement than mExos@DSPC), whereas stiff mExos@DSPC hybrids were internalized 2.5-fold more by Caco-2 epithelial cells, primarily via caveolin-mediated endocytosis, and were subsequently routed through the ER–Golgi secretory pathway rather than toward lysosomes. Transepithelial permeability of mExos@DSPC was 2.9-fold higher than mExos@DMPC, translating to a relative oral bioavailability of 5.9% versus 3.1% for semaglutide and superior glycemic control in a T2DM rat model over six weeks [45]. Collectively, these studies demonstrate that post-biogenesis modifications such as cholesterol enrichment, polymer grafting, and phospholipid hybridization, can redirect the dominant cellular entry pathways, extend circulation time, and improve in vivo therapeutic efficacy. The cholesterol-loading and hybridization studies establish a direct causal link between membrane stiffness and functional outcome, whereas the contribution of mechanical changes in PEGylation and polymer-grafting approaches for EVs remains to be studied. Disentangling the mechanical versus chemical contributions of surface modifications will require a systematic nanomechanical characterization with functional assays across a broader range of formulations. Culture conditions as an upstream mechanical lever Beyond cell source and post-isolation modification of EVs, the mechanical microenvironment of the producer cell offers an additional lever to engineer the yield, molecular cargo, and the function of the EVs. Substrate stiffness is the most extensively studied mechanical input. In tumor cells, stiff extracellular matrix (ECM) activates integrin-FAK-Akt signaling, which phosphorylates the guanine nucleotide exchange factor Rabin8 and promotes GTP loading, ultimately driving EV secretion [95]. l This finding links substrate rigidity to tumor-promoting EV subpopulations through a defined mechanotransduction cascade, although the mechanical properties of the EVs themselves were not characterized in the study. In MSCs, softer ECM substrates promote higher EV secretion per cell, while stiffer substrates drive preferential sorting of vesicular-transport proteins and autophagy-related lipids into EVs, resulting in enhanced downstream macrophage uptake [98]. Together, these two studies demonstrate that substrate stiffness programs EV composition through cell-type-specific mechanotransduction pathways and affects their downstream fate, although the generality of these pathways across other cell types remains to be tested. 3D culture geometry provides further control of EV production. Spheroid confinement of MSCs increased EV output relative to monolayer culture, attributed to the transition of cells to a non-adherent, round morphology [99]. However, this 3D culture geometry does not guarantee higher yield or improved function. Kusuma et al. compared MSC-EVs from 2D monolayer and 3D spheroid cultures and found no significant difference in EV yield, but proteomic profiling revealed cargo difference, with 3D EVs were enriched in immune- and ECM/fibrosis-related proteins. This cargo shift did not translate into functional benefit, with 3D EVs showing reduced indoleamine 2,3-dioxygenase (IDO) activity and reduced macrophage phagocytosis, and in a bleomycin-induced pulmonary fibrosis mouse model exacerbated disease by increasing collagen deposition and worsening respiratory mechanics compared with 2D EVs [100]. These results highlight that while culture geometry reprograms EV cargo, such alterations can be detrimental, making functional validation crucial when modifying production protocols. Dynamic bioreactor environments introduce fluid shear as an additional mechanical input. Jeske et al. showed that vertical-wheel bioreactor expansion of MSCs on microcarrier substrates increased EV yield 2.5-fold relative to static 2D culture, with upregulation of both ESCRT-dependent and ESCRT-independent biogenesis genes and remodeling of miRNA and protein cargo toward metabolic, autophagy, and ROS-related profiles [101]. However, this transition from 2D to dynamic 3D culture involves both microcarrier-based 3D geometry and fluid shear simultaneously, and the experiment design does not permit the separation of their individual contributions. Together, it has been shown that the upstream mechanical inputs of the parent cells, including substrate stiffness, culture dimension, and shear exposure, can be used to tune EV yield and cargo composition as scalable manufacturing parameters. However, two significant gaps remain. First, the mechanotransduction pathways linking the mechanical culture condition to EV biogenesis have been characterized only in a few cell types, and a deeper mechanistic understanding of how specific mechanical inputs alter cargo loading would enable more rational design of culture conditions. Second, whether these upstream mechanical inputs also alter the biophysical properties of the secreted vesicles, such as membrane stiffness, bending modulus, or deformability, has not been directly tested. Given the evidence from Sect. 3.3.1 that EV mechanics influence biological barrier traversal and cellular uptake, systematic nanomechanical characterization of EVs produced under defined culture conditions would clarify whether culture mechanics act solely through cargo programming or also through mechanical specification of the vesicle itself. Future perspectives and challenges Although the mechanical regulation of LNCs is increasingly recognized as a pivotal determinant of their biological fate, translating these biophysical principles into clinical therapeutics remains a challenge. Overcoming existing bottlenecks requires a synergistic, interdisciplinary approach that bridges materials engineering, advanced metrology, and computational biology. Standardization of nanomechanical measurements Achieving greater standardization of nanomechanical measurements is widely recognized as an important challenge that has yet to be fully resolved. The survey of the existing literature revealed that critical AFM methodological details, such as instruments, substrate preparation, cantilever spring constant calibration method, contact mechanics model selection, indentation depth, buffer composition, and measurement temperature, were not consistently reported across studies, which further hampers cross-study comparisons. Full reporting of experimental parameters alongside deposition of raw force-indentation curves is therefore essential to ensure data reproducibility and inter-laboratory compatibility. Beyond reporting, the field requires high-throughput, in situ characterization platforms such as real-time deformability cytometry and optical tweezers. Coupling these technologies with standardized protocols will help to resolve intra-batch heterogeneity, allowing researchers to probe mechanics in physiological environments and build a reliable foundation for clinical evaluation. Development of comparable datasets Establishing a comparable database that systematically links vesicle mechanics to biological performance would represent a meaningful step forward. Standardized protocols form the foundation for constructing comparable, open-access datasets; however, this objective is currently impeded by the structural and mechanical heterogeneity inherent in scalable manufacturing. Intra-batch variations in mechanical properties can drive inconsistent cellular interactions and unpredictable pharmacokinetic outcomes. Such datasets must link nanomechanical measurements to compositional variables and biological performance outcomes, forming the structure-mechanics-biology triplets that give the database genuine predictive value. Integration of mechanics into formulation optimization A clearer understanding of how mechanical properties influence the nanocarrier biological journey, such as in vivo circulation, cellular uptake, endosomal escape, and immune evasion, could offer an additional, complementary handle for rationally tuning delivery performance alongside established parameters such as size and surface chemistry. How to meaningfully incorporate these mechanical insights into formulation optimization, rather than treating them as secondary descriptors measured after compositional decisions have already been made, remains an open and important question. One promising route is to close the feedback loop between comparable datasets and AI-driven formulation models, where predicted mechanical consequences of compositional changes are evaluated iteratively alongside experimental results, gradually shifting development from empirical screening toward more guided design. Realizing this vision, however, first requires the mechanistic datasets and standardized measurement infrastructure described above, without which mechanics-informed formulation optimization remains aspirational rather than practical. Conclusion The mechanical properties of LNCs have moved beyond passive structural consequences of nanocarrier formulation. The studies reviewed here establish that the mechanical properties of LNCs are now tunable parameters with measurable biological consequences. Mechanics thus provides a link between synthetic formulation design and in vivo performance. However, mechanical characterization of LNCs is still not routinely performed across the field, and a systematic understanding of how specific mechanical parameters govern biological outcomes has yet to be established. Standardized, high-throughput measurements and systematic correlation of mechanical properties with biological consequences will help transition from empirical trial-and-error to mechanics-guided rational design. Incorporating mechanics as an additional design axis, alongside other established parameters such as composition and surface chemistry, provides a new dimension for addressing the translational challenges. Addressing these metrological challenges and integrating predictive modeling would help establish mechanics as a routine, quantitative axis of nanocarrier design alongside composition and surface chemistry.