In Vitro Models for Peptide Evaluation: Cell Lines, Organoids, and Core Assays

Cell line choice constrains everything downstream. For dermal research, HaCaT —a spontaneously immortalized human keratinocyte line— has become the workhorse: it expresses epidermal differentiation markers, secretes cytokines such as IL-1α and IL-6, and responds reproducibly to inflammatory stimuli and bioactive peptides. For hepatocytes, HepG2 (human hepatocellular carcinoma) remains the default for early toxicity and metabolism screens, though its metabolic capacity is reduced relative to primary human hepatocytes.

In skeletal muscle, C2C12 mouse myoblasts allow studying both proliferation and differentiation into myotubes, useful for peptides with anabolic signalling interest. HUVEC (human umbilical vein endothelial cells) cover the vascular front: angiogenesis, endothelial migration, permeability. For neuroscience, differentiated SH-SY5Y and PC12 are the typical entry points, even if neither truly mimics adult human neurons.

Working rule: the line should match the tissue where activity is expected, not the line that happens to be in the incubator. A peptide active in HepG2 tells you little about skin; a peptide that drives HaCaT migration tells you little about endothelium.

Immortalized lines (HaCaT, HepG2, C2C12) are reproducible across batches, grow fast, and are inexpensive. That same stability is their weakness: they come from a genome that went through an immortalization crisis, with altered karyotypes and, often, blunted stress-response pathways. Primary cultures —normal human epidermal keratinocytes (NHEK), primary human hepatocytes (PHH), muscle satellite cells— preserve donor-tissue phenotype better, but introduce inter-donor variability, short useful life, and a much higher per-experiment cost.

Best practice in 2026 for serious programs is to use immortalized lines for primary dose-response screening, then validate hits in primaries or 3D systems. Jumping straight to primaries for exploratory screening burns budget without proportional information; reporting results only in an immortalized line without follow-up validation is hard to defend.

2D culture ignores oxygen gradients, extracellular matrix interactions, and tissue architecture. 3D organoids —spheroids, stem-cell-derived organoids, reconstructed skin equivalents— recover part of that complexity. For dermal research, reconstructed human epidermis built from HaCaT or primary keratinocytes on collagen matrices has been used for years as an irritation and percutaneous penetration model.

Advantages are clear: higher physiological relevance, extended viability (spheroids hold for up to three weeks, versus less than one week for a confluent monolayer), gradients that recapitulate in vivo microenvironment. Limitations are also real: higher cost, lower throughput, more batch-to-batch variability, and lack of vasculature and dynamic flow unless coupled with microfluidics (organ-on-chip).

Operational take: do not migrate the whole pipeline to 3D. Keep 2D for library screening and initial dose-response; reserve 3D to confirm hits, to do penetration studies, and for experiments where tissue architecture is central to the question.

Viability and proliferation are covered by tetrazolium salt assays. MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) produces an insoluble formazan that needs isopropanol or DMSO to dissolve; WST-1 yields a water-soluble formazan directly in medium, avoiding the lysis step and reducing noise. In modern peptide work, WST-1 or CCK-8 tend to be preferred for lower reagent cytotoxicity. Resazurin (AlamarBlue) allows repeated measurements on the same wells across time.

For cell migration, the scratch assay remains the most popular for cost and simplicity: a pipette tip wounds a confluent monolayer, photographs are taken at t=0, and closure is measured at 6/12/24 h. To separate migration from proliferation it is important to add mitomycin C (DNA synthesis inhibitor) or to work in serum-free medium. Tighter alternatives: Boyden chambers, transwell systems, and removable-barrier platforms.

Gene expression is measured by qPCR for targeted hits, or RNA-seq when a broad profile is needed. For secreted proteins —cytokines, growth factors, matrix markers— ELISA is still the standard; multiplex arrays (Luminex) let you measure dozens of analytes from a single supernatant. Choice scales with how focused the hypothesis is.

A positive in vitro result is necessary, never sufficient. Three frequent traps. First, concentrations: testing a peptide at 100 µM in culture and reporting it as active says little about concentrations reachable in plasma or tissue in vivo, typically one or two orders of magnitude lower. Second, model: a pro-migratory effect in HaCaT does not guarantee skin healing in humans; immunity, vasculature, tissue mechanics, and the microbiome are all absent.

Third, specificity: many short peptides produce nonspecific effects at high concentrations (membrane interaction, aggregation, synthesis impurities). Proper controls —scrambled peptide, length-matched non-active peptides, oxidized or reduced peptide where relevant— separate real signal from artifact. Without them, the readout is diluted.

Every in vitro observation discussed here should be read as exploratory preclinical evidence. It is not a basis for human use, nor for dose extrapolation. The model's value lies in generating mechanistic hypotheses, prioritizing candidates, and ruling out toxics before committing resources to in vivo work.