Covalent organic frameworks have emerged as serious contenders in the porous materials arena, and porphyrin-based variants are particularly interesting because they marry crystalline order with photophysical muscle. The catch has always been size. Bulk COFs sit on the shelf beautifully but behave poorly in biological environments. Cells do not welcome giant flakes; tissues block them. What the field needed was a way to shrink these frameworks without destroying their internal architecture, and then to hand them a biological GPS. A recent report in International Journal of Biological Macromolecules tackles exactly this challenge by splitting a single porphyrin COF into two distinct size classes, wrapping each in hyaluronic acid, and deploying them in a choreographed sequence that first remodels the extracellular space and then breaches the plasma membrane.

The synthesis itself is a textbook solvothermal Schiff-base reaction between tetraformylphenyl porphyrin and 4,4′-diaminodiphenyl disulfide. Acetic acid catalysis, three freeze-thaw cycles, and three days at 120 °C yield the bulk material. The clever step comes next: ultrasonic homogenization at different intensities fractures the bulk into large P-COF plates around 500 nm and smaller p-COF fragments near 200 nm. Both retain their crystallinity, as powder X-ray diffraction confirms, and both show the expected imine stretch at 1596 cm⁻¹ in the infrared. Solid-state NMR locks down the 159.6 ppm carbon signal that proves the C=N linkage is real. Nitrogen sorption gives surface areas above 240 m²/g with a tight 3.9 nm pore distribution, so the porphyrin units remain accessible to oxygen and light. Even after aggressive sonication, the smaller p-COF does not turn amorphous, which is critical because once the crystal structure collapses, the photosensitizer aggregates and quenches its own excited states.

Hyaluronic acid coating is the next layer of sophistication. HA does not just stabilize colloids; it creates a viscous surface mesh that shifts zeta potential sharply negative and provides a docking mechanism for CD44-rich surfaces. TGA quantifies the loading: about 120 µg HA per mg of P-COF@HA and roughly 840 µg per mg of p-COF@HA. The difference makes sense because the smaller particles present more surface area for the polymer to grip. In water, RPMI 1640, and DMEM, neither size class aggregates over a week, which is a non-negotiable prerequisite for anything headed into a biological setting. The mechanism is physical adsorption, aided by electrostatic complementarity between the negatively charged HA and the COF surface.

The photochemistry is where porphyrin COFs earn their keep. Under 660 nm irradiation, the framework photosensitizers jump to an excited triplet state and dump energy into surrounding oxygen, generating singlet oxygen. DPBF bleaching assays and EPR spin-trapping with TEMP both confirm the characteristic ¹O₂ signature. Interestingly, the smaller p-COF outperforms its larger sibling, likely because the reduced dimensions cut down on self-quenching and improve oxygen access to the porphyrin cores. Even after partial disulfide cleavage by reduced glutathione, the fragments still produce enough reactive species to remain photophysically relevant. This is a practical point often overlooked: many responsive nanocarriers fall apart so thoroughly that they lose their primary function. Here, the COF shrinks but keeps singing.

This GSH responsiveness is not a side effect; it is a design feature. The disulfide bridges woven into the COF backbone are redox-active. In environments rich in thiols, they cleave, fracturing the framework and releasing whatever cargo has been loaded. DTNB assays show a time- and pH-dependent consumption of GSH: the more acidic the medium and the longer the incubation, the more disulfide bonds break. TEM images track the morphological collapse as GSH concentration climbs from 0 to 20 mM. The imine bonds add a second layer of acid sensitivity, so the construct effectively senses two chemical variables at once. By stripping GSH from the local environment, the COF also removes one of the cell’s primary antioxidant defenses, which indirectly amplifies any subsequent oxidative insult generated by the porphyrin core under illumination.

The researchers loaded the large P-COF with losartan, an agent known to down-regulate collagen I and α-SMA expression, thereby softening the extracellular matrix and reducing solid stress. The smaller p-COF was loaded with doxorubicin, a DNA-intercalating compound. Both payloads hitch a ride via π-π stacking and, in the case of the positively charged doxorubicin, electrostatic attraction to the negatively charged COF surface. Release profiles are telling: at neutral pH, doxorubicin leakage stays below 35 % over 36 hours, but in acidic, GSH-rich, hyaluronidase-present conditions, cumulative release climbs past 75 %. Losartan shows a similar pH-biased release pattern. The system is effectively locked in transit and unlocked at the destination.

Cellular uptake studies on 4T1 cells reveal the spatial logic of the dual-size approach. Because CD44 is abundantly expressed on these cells, the HA-coated particles show strong binding. The large LCH particles adhere to the outer membrane rather than entering, which is exactly what they are supposed to do. Under laser exposure, they generate ¹O₂ at the membrane surface, triggering lipid peroxidation detected by BODIPY-C11 fluorescence. This oxidative damage increases membrane permeability and instability, priming the cell for subsequent events. The mechanism is straightforward: reactive species attack polyunsaturated lipids, generate lipid radicals, and create membrane defects that lower the energy barrier for nanoparticle entry. Flow cytometry and confocal imaging confirm that the smaller DCH particles penetrate efficiently, especially when the membrane has been pre-conditioned. GSH levels inside the cells drop dose-dependently after p-COF@HA exposure, confirming that the disulfide chemistry is actively stripping intracellular thiols and simultaneously setting the stage for enhanced oxidative stress.

In three-dimensional multicellular spheroids, the sequential strategy shows its full value. LCH pre-administration degrades the collagen-rich extracellular barrier, while laser-induced lipid peroxidation punches holes in the membrane. When DCH is introduced afterward, it penetrates deeper into the spheroid core than it ever could alone. ROS production, measured by DCFH-DA, is strongest in the sequential LCH-plus-DCH group, and spheroid growth suppression follows the same hierarchy. The concentration needed to halve proliferation drops by roughly half compared to free compounds, largely because the nanoformulation bypasses efflux pumps and enters through endocytic routes that free molecules cannot exploit.

To map the in vivo fate of the nanosystem, the team required a reliable optical tracer. They selected Cy5, a near-infrared fluorescent dye obtained from AbMole, and conjugated it to the DCH platform. At an administered concentration of 2 mg/kg via intravenous administration, DCH-Cy5 produced a robust signal that peaked around 36 hours post-administration. Ex vivo imaging at 48 hours showed pronounced fluorescence retained at the target site, with negligible accumulation in the heart, liver, spleen, lung, and kidney. The AbMole Cy5 tracer effectively demonstrated that the HA-directed CD44 targeting successfully concentrated the payload where intended, validating the design logic in a living system without ambiguous background noise. Clean biodistribution data are surprisingly hard to come by in nanoparticle research; many particles linger in the reticuloendothelial system and create false positives. The Cy5 signal here was crisp and site-specific.

Transcriptome sequencing of samples from the sequentially administered cohort versus controls identified 889 differentially expressed genes. ECM-related transcripts including Col20a1, Col8a1, Cspg5, and Fn1 were downregulated, while Adamts5 moved in the opposite direction. GO and KEGG analyses pointed to extracellular matrix organization, collagen-rich scaffold modulation, and cell-surface receptor signaling as the primary affected categories. The data suggest that the platform does not merely deliver compounds; it rewrites local gene expression to favor a more permissive microenvironment. Notably, Smad3, a central mediator of matrix production, was also suppressed, which aligns with the observed drop in collagen I and α-SMA protein levels.

Biocompatibility data back up the materials side. Normal hepatocyte LO2 cells tolerate p-COF concentrations up to 200 µg/mL with survival rates above 96 %. Hemolysis stays below 0.7 % even at the same concentration, and blood chemistry markers remain unperturbed. The selectivity arises because normal cells do not harbor the high GSH levels needed to rapidly disassemble the disulfide-linked framework, so the cargo stays put and the carrier remains intact. This creates a useful selectivity window: the chemistry is the same everywhere, but the response is tuned by the local redox potential.

What stands out about this study is the refusal to treat biological barriers as a single problem to be solved with a single particle. The extracellular matrix and the plasma membrane are fundamentally different obstacles, and they are addressed here by fundamentally different tools: a large, membrane-anchored COF that softens the outer fortress, and a small, intracellular COF that executes the internal program. The hyaluronic acid coating unifies them under one targeting philosophy, while the porphyrin core provides the light-activated engine. By tuning size instead of chemistry, the researchers get two distinct behaviors from one synthetic parent. It is an elegant reminder that in nanomaterials, geometry is often as powerful as composition. Future work will likely explore whether this sequential logic can be extended to other size-sensitive barriers, such as the blood-brain barrier or mucosal layers. For now, the platform offers a compelling proof that intelligent timing and dimensional control can turn a simple porphyrin framework into a sophisticated, multi-stage delivery machine.