LX-2 hepatic stellate cells alter 3D HepG2 hepatocellular carcinoma spheroids and are associated with lower sorafenib sensitivity

In an in vitro cell-line study, researchers found that incorporating LX-2 hepatic stellate cells into three-dimensional HepG2 hepatocellular carcinoma spheroids was associated with more compact architecture, altered cell-death and metabolic measurements, and reduced sensitivity to sorafenib. The investigators reported an estimated concentration required for 30% inhibition of 25.9 micromolar in coculture spheroids, compared with 8.5 micromolar in HepG2 monocultures.

The experiment compared three spheroid models: HepG2 hepatocellular carcinoma cells alone, LX-2 hepatic stellate cells alone, and a 1:1 coculture containing both cell types. Each well received a total of 50,000 cells, with cocultures containing 25,000 cells of each type. Spheroids formed through spontaneous self-assembly in agar-coated 96-well U-bottom plates and were followed for 96 hours. The study combined longitudinal phase-contrast imaging and ImageJ-based morphometry with scanning electron microscopy, flow cytometry, biochemical assays, and a CellTiter-Glo 3D viability assay.

HepG2 monoculture spheroids had a projected area 4.6 times greater than coculture spheroids and 4.8 times greater than LX-2 spheroids at 96 hours, with both comparisons reported as statistically significant. Projected area in LX-2 and coculture spheroids declined by 72% and 71.7%, respectively, during the first 24 hours. By 96 hours, area had fallen 80.42% in LX-2 spheroids and 77.9% in coculture spheroids, while the 21.9% decline in HepG2 spheroids was not statistically significant.

Other shape measurements also distinguished the models. At 96 hours, mean circularity was 0.1373 for HepG2 spheroids, 0.4045 for LX-2 spheroids, and 0.4112 for cocultures; HepG2 values were significantly lower than those of both comparison models. Perimeter reductions from 2 to 96 hours were 77.0% for LX-2 and 70.8% for coculture spheroids, compared with a nonsignificant 23.8% reduction for HepG2. An image-derived optical compaction index also rose significantly over time in LX-2 and coculture spheroids. The researchers cautioned that this index was an indirect measure of apparent optical compaction, not a direct measurement of cellular density, extracellular matrix content, or mechanical stiffness.

Scanning electron microscopy after 96 hours showed a compact, continuous surface on LX-2 spheroids and a rough, porous surface on HepG2 spheroids. Coculture spheroids had an intermediate appearance, with a more continuous external organization than HepG2 spheroids but remaining surface irregularities. These observations were consistent with the quantitative morphometry, although the study did not directly measure extracellular matrix mechanics or establish which processes caused the structural differences.

Cell-death measurements were broadly similar for viable, early-apoptotic, and late-apoptotic populations. Coculture spheroids had a lower mean proportion of propidium-iodide-positive cells than HepG2 spheroids, but that pairwise comparison did not reach statistical significance. The percentage of active-caspase-3-positive cells did not differ significantly among models. Fluorescence intensity did, however, because cocultures had lower active-caspase-3 signal than HepG2 spheroids. When cells in cocultures were operationally separated by hepatocyte nuclear factor 4 alpha expression, the HepG2-like population had approximately 28% lower signal than HepG2 monocultures, while the putative LX-2-like population had approximately 76% lower signal than LX-2 monocultures. Because caspase-3 activation is transient, the fluorescence measurements do not by themselves establish a lower rate of apoptosis.

Biochemical testing found that intracellular LDH activity did not differ significantly among models, while extracellular LDH activity was highest in HepG2 spheroids, intermediate in cocultures, and lowest in LX-2 spheroids. LDH release normalized to total measurable LDH did not differ significantly between HepG2 and coculture spheroids, so the normalized result did not independently confirm better membrane preservation in cocultures. Extracellular lactate concentrations in conditioned medium were highest in LX-2 spheroids, intermediate in cocultures, and lowest in HepG2 spheroids. These values were not normalized to cell number, DNA, protein content, or biomass and therefore represented concentrations in the medium rather than cell-specific lactate production rates.

Flow cytometry also identified stromal-associated matrix signals. Approximately 24% of coculture singlet cells were alpha-smooth muscle actin-positive after 96 hours, although marker-defined fractions did not provide an exact measurement of cellular composition. Collagen type I alpha 1-associated fluorescence was higher in cocultures than in HepG2 monocultures and was concentrated in the alpha-smooth muscle actin-positive compartment. This supported the presence of a collagen-producing, activated-like stromal phenotype, but the experiment did not directly quantify collagen deposition, extracellular matrix organization, or tissue stiffness.

Coculture spheroids retained greater viability than HepG2 spheroids at 15 and 25 micromolar sorafenib, and concentration-response analysis produced an estimated IC30 approximately three times higher than that of HepG2 monocultures. The findings associate LX-2 inclusion with lower sorafenib sensitivity rather than complete resistance, because coculture viability declined at the highest concentrations tested. The study did not measure intracellular drug accumulation, penetration, matrix stiffness, or specific signaling pathways, so it could not determine whether reduced sensitivity arose from structural barriers, paracrine signals, or both.

The principal limitation was the use of established HepG2 and LX-2 cell lines, which do not reproduce the genetic, phenotypic, or interpatient diversity of primary tumors and stromal populations. LX-2 cells also have a constitutively activated-like phenotype and cannot represent the full range of hepatic stellate cell states found in human liver tumors. The results therefore characterize this defined in vitro system rather than clinical sorafenib response in people with hepatocellular carcinoma.

Subject of Research: Effects of LX-2 hepatic stellate cells on 3D HepG2 spheroid architecture, biomarkers, metabolism, and sorafenib sensitivity

Article Title: Hepatic stellate cells promote structural and functional remodeling of three-dimensional hepatocellular carcinoma coculture spheroids

Article References: Obelar, J., Vasconcellos, J. G., do Nascimento, N. B., de Sousa, A. C., Giehl, M., Simões Fischer, B., de Souza Ferreira, J., de Oliveira, J., Trindade, V. M. T., & Guma, F. T. C. R. (2026). Hepatic stellate cells promote structural and functional remodeling of three-dimensional hepatocellular carcinoma coculture spheroids. Molecular Biology Reports, 53(1), Article 1566. https://doi.org/10.1007/s11033-026-12715-9

Image Credits: AI Generated

DOI: 10.1007/s11033-026-12715-9

Keywords: hepatocellular carcinoma, hepatic stellate cells, three-dimensional culture, spheroids, extracellular matrix, COL1A1, sorafenib, cell death, tumor microenvironment, Hepatic, stellate, cells