Human serum albumin (HSA) is a workhorse carrier for nanomedicine: it is abundant, biocompatible, and engages two tumour-relevant receptors, gp60 and SPARC. This article walks through a complete experiment — from designing a self-assembling HSA variant with a reversed QTY code, to preparing nanomicelles, to a full characterisation pipeline spanning turbidimetry, cell viability, endocytosis inhibition, and orthotopic tumour studies. The thread connecting the sections is that every design choice has to be re-verified by an assay: the protein design predicts behaviour, but turbidity, viability, and uptake decide whether it is true.
Protein design: the reversed QTY code
The original QTY code converts hydrophobic α-helical residues into hydrophilic ones (L→Q, V→T, I→N, F→Y), producing water-soluble variants of membrane proteins that otherwise need detergent. The strategy here runs the logic in reverse. In three α-helical subdomains of natural HSA — IB, IIB and IIIB — hydrophilic residues (Q/N, T and Y) are partially replaced by hydrophobic residues (L, V and F), respectively. The substitutions are chosen on the basis of similarity in the electron-density maps of the amino-acid side chains, so the mutated side chains pack into the helix with a shape and volume close to the original, while the surface character flips from polar to apolar. The result, HSArQTY, keeps the HSA fold but becomes amphipathic enough to self-assemble in aqueous buffer — an "inside-out" use of the same code, where the goal is not solubility but controlled self-association into a micellar carrier.
The design step is dry-lab work, but it constrains everything downstream: the number and position of the reversed substitutions set the amphipathicity, and therefore the particle size and the critical micelle behaviour you will later measure.
Preparation of HSArQTY nanomicelles
The nanomicelles are made by a solvent-injection/sonication method, which is the practical route to self-assembled protein micelles from a soluble building block:
- Dissolve purified HSArQTY in a small volume of aqueous buffer or a water-miscible solvent to give the stock "protein solution." Clarify by centrifugation and filter (0.22 µm) before use.
- Bring PBS to the standing volume, place it under the ultrasonic probe, and start sonication. Sonication supplies the shear energy that disperses the injected protein as it contacts the aqueous phase.
- Inject the HSArQTY solution into the PBS under continuous sonication. The injection should be steady and dropwise, not a single bolus; abrupt addition gives large, polydisperse aggregates instead of uniform nanomicelles.
- When the injection is complete, stop the ultrasonic dispersion immediately. Prolonged sonication after assembly can re-disperse or fragment the micelles and heat the sample.
- Recover the nano-micelle suspension. Characterise size and polydispersity by dynamic light scattering (DLS) and verify morphology by electron microscopy before any biological use.
Practical notes. Keep the probe tip submerged but away from the vessel wall to avoid metal shedding; work in an ice bath to limit heating; and record power, pulse regime, and total sonication time, since micelle size is sensitive to all three. If DLS shows a high polydispersity index or a second population, the usual causes are too-fast injection, an underpowered probe, or protein aggregation in the stock.
Turbidity measurements
Micelle formation and colloidal stability can be followed by turbidity, measured as nephelos turbidity units on a calibrated portable nephelometer. Turbidity rises as self-assembled particles scatter light, so a stable, low and reproducible reading indicates a well-dispersed suspension, while a rising or drifting value signals aggregation or precipitation. Because the reading depends on the instrument's optics and calibration, always: (i) calibrate with the supplied standards before the run; (ii) measure blank buffer at the same temperature; and (iii) use matched, scratch-free cuvettes, since dust and fingerprints dominate the signal at low turbidity. Report the unit, temperature, and any dilution.
Cell viability: CCK-8 and live/dead staining
Biocompatibility is the gating safety assay for any carrier. Two readouts are used here across two cell lines — HEK293T (human embryonic kidney) and L929 (mouse fibroblast):
- CCK-8: seed cells in 96-well plates, allow attachment, then expose to a dilution series of the three HSArQTY nanoparticle preparations alongside untreated and vehicle controls. After the incubation, add CCK-8 reagent (WST-8), which is reduced by viable cells to a water-soluble formazan, and read absorbance at ~450 nm. Convert to percent viability against the untreated control; report IC50 only if a dose–response actually reaches it.
- Live/dead staining: after co-culture with the nanoparticles, stain with a live/dead cell staining kit (calcein-AM for live cells, a membrane-impermeant nucleic-acid dye such as propidium iodide or ethidium homodimer for dead cells) and image by fluorescence microscopy. Co-register the live and dead channels to distinguish genuinely dead cells from detached or debris signal.
Run both assays in at least triplicate with independent preparations; a single batch can hide batch-to-batch variation in self-assembly.
Endocytosis-inhibition experiments
To ask which uptake route the nanoparticles use, two tumour cell lines (4T1, A549) and two normal cell lines (HEK293T, L929) are compared:
- Profile the expression of the albumin receptors gp60 and SPARC by immunofluorescence. Higher receptor density in tumour cells is the mechanistic hypothesis being tested, so this step establishes the premise before any inhibition.
- Block endocytosis in the tumour cells by adding chlorpromazine (clathrin-mediated pathway), amiloride (macropinocytosis), and β-cyclodextrin simultaneously, then measure how much uptake is lost. Using the three inhibitors together probes the major routes at once; follow up with single inhibitors if you need to attribute uptake to one pathway.
- In HUVECs, add methyl-β-cyclodextrin to disrupt the gp60 receptor function and test the gp60-dependent arm specifically. Cholesterol-extraction by methyl-β-cyclodextrin perturbs the membrane environment that gp60-mediated transcytosis requires.
Include vehicle-only controls for every inhibitor, because these compounds can themselves reduce viability; normalise uptake to viability when the treatment is long.
Animal experiments: orthotopic 4T1 tumour model
The in vivo work uses an orthotopic, bioluminescent breast tumour model:
- Inject 1 × 107 4T1-luc cells subcutaneously into the right forelimb of each mouse. The luciferase reporter allows tumour burden to be followed non-invasively.
- Allow tumour-bearing mice to establish — approximately 5 days — then randomise into treatment groups and dose according to the group design.
- Record survival and draw the percent survival curve for each group from the recorded data (Kaplan–Meier style, with censoring noted).
- For tumour-targeting analysis, inject IR783-labelled nanoparticles into the caudal vein and image the near-infrared signal over time to follow accumulation at the tumour.
- For in vivo observation of tumour size, inject D-luciferin potassium working solution and acquire bioluminescence images; quantify total flux as a surrogate for viable tumour mass.
Limitations and controls
- Bioluminescence caveats: luciferase expression can impose a metabolic burden and does not report necrotic tissue, so bioluminescence and caliper measurements should be interpreted together.
- Inhibitor specificity: chlorpromazine, amiloride, methyl-β-cyclodextrin and β-cyclodextrin are not perfectly selective; treat the inhibition profile as evidence for a dominant route, not proof of exclusivity.
- Dye leakage: IR783 can dissociate from the carrier, so include free-dye controls to separate true targeting from free-probe biodistribution.
- Batch reproducibility: nanomicelle size is sensitive to the preparation, so characterise every batch by DLS and turbidity before pooling biological results.
References
- Zhang S, Tao F, Qing R, et al. QTY code enables design of detergent-free chemokine receptors that retain ligand-binding activities. Proc Natl Acad Sci USA. 2018;115(37):E8652–E8659. doi:10.1073/pnas.1811031115
- Jahanban-Esfahlan A, Dastmalchi S, Davaran S. A simple improved desolvation method for the rapid preparation of albumin nanoparticles. Int J Biol Macromol. 2016;91:703–709. doi:10.1016/j.ijbiomac.2016.05.032
- Abbasi S, Paul A, Shao W, et al. Cationic albumin nanoparticles for enhanced drug delivery to treat breast cancer: preparation and in vitro assessment. J Drug Deliv. 2012;2012:686108. doi:10.1155/2012/686108
- Desai N. Nanoparticle albumin bound (nab) technology: targeting tumors through the endothelial gp60 receptor and SPARC. Nanomedicine. 2007;3(4):339. doi:10.1016/j.nano.2007.10.021
- Schnitzer JE, Oh P. Antibodies to SPARC inhibit albumin binding to SPARC, gp60, and microvascular endothelium. Am J Physiol. 1992;263(6):H1872–H1879. doi:10.1152/ajpheart.1992.263.6.h1872
- Ji Q, Zhu H, Qin Y, et al. GP60 and SPARC as albumin receptors: key targeted sites for the delivery of antitumor drugs. Front Pharmacol. 2024;15:1329636. doi:10.3389/fphar.2024.1329636
- Baklaushev VP, Kilpeläinen A, Petkov S, et al. Luciferase expression allows bioluminescence imaging but imposes limitations on the orthotopic mouse (4T1) model of breast cancer. Sci Rep. 2017;7:7715. doi:10.1038/s41598-017-07851-z