Engineered Huh7-based microtissues showed a more liver-like pattern across gene, protein and metabolic tests when they carried an HNF4α-enhancing RNA tool, the laboratory study reports. Compared with control tissues made from parental, non-transfected Huh7 cells, the engineered tissues had higher liver-associated transcripts and proteins, secreted more ALB and less AFP, released less lactate and showed more glycogen staining. The report also recorded higher activity in its inducible CYP assays, which were used to examine drug-processing function.
The finding comes from a three-dimensional co-culture model, meaning different cell types were grown together in a compact tissue. It combined Huh7 cells with human umbilical vein endothelial cells and Wharton's jelly mesenchymal stem cells, alongside microparticles made from sheep liver extracellular matrix. Experimental tissues used the HNF4α-enhanced Huh7 cells, while controls used parental, non-transfected Huh7 cells under the same culture conditions.
How the tissues were built
To create the experimental Huh7 cells, the researchers stably introduced a SINEUP-based antisense long noncoding RNA, a genetic tool intended to increase HNF4α production. They selected the cells with 700 μg/mL G418 starting 72 hours after transfection for approximately seven days, then used qPCR to verify expression.
Each microtissue used Huh7, HUVEC and WJ-MSC cells at a 10:7:2 ratio, with two parts cells for every one part of the liver-matrix microparticles. The microparticles were mostly spherical and usually 3 to 7 micrometers across, with no significant viability reduction in the tested Huh7 or HFF cells. Both experimental and control groups formed compact, uniform microtissues with comparable size distributions and predominantly viable cells.
The molecular profile looked more liver-like
At the transcript level, the engineered tissues had more HNF4α and ALB than controls and less AFP. HNF4α and ALB were described as approaching the healthy-liver reference. The reported differences met thresholds of p < 0.05 for HNF4α, p < 0.01 for ALB and p < 0.001 for AFP.
The shift extended beyond differentiation markers. CDH1 was higher and CDH2 lower, producing an epithelial-leaning cadherin pattern that resembled normal liver. The glycolysis-related markers HK2, LDHA and GAPDH fell, while the gluconeogenesis-related markers G6PC, PCK1 and GCK rose. APOA1 and APOB increased, while SOAT1 and SOAT2 decreased.
The protein tests broadly matched the RNA findings. Immunofluorescence showed more HNF4α, CYP3A4, E-cadherin and GLUT2, and less N-cadherin and GLUT1 in engineered tissues. Western blotting found higher HNF4α, E-cadherin and GLUT2 and lower N-cadherin, GLUT1 and GAPDH. The six measured CYP transcripts, gene-level readouts used in the drug-processing tests, were also higher than in controls and were described as approaching or approximating the liver reference.
Function was stronger in some tests, mixed in others
Secretions and other functional assays gave a generally more specialized profile. Engineered tissues secreted more ALB and less AFP than controls, while fibrinogen production was unchanged. They released less lactate, produced more urea and showed more intense glycogen staining. In a 72-hour migration assessment, they spread radially less and showed less structural dissociation.
The drug-response pattern varied by compound. After 48 hours, viability was lower in engineered tissues than in controls at APAP doses of 10 and 100 mM and INH doses of 1,000 and 5,000 μM. After 5-FU exposure, however, viability was higher in the engineered group.
In separate inducible CYP tests, the report said the engineered tissues had higher activity than controls for CYP3A4, CYP2B6, CYP2C9 and CYP1A2. Phenobarbital was used for the CYP3A4 and CYP2B6 tests, rifampicin for CYP2C9, and omeprazole for CYP1A2.
A useful signal, with important limits
The authors interpret the findings as partial metabolic improvement and an incomplete shift toward a hepatocyte-like phenotype. They do not present the tissues as a full restoration of hepatocyte function. Instead, they propose the system as a complementary, scalable in vitro platform for drug screening and say that drug-metabolism and CYP findings need further validation.
The result also rests on a small number of biological comparisons. Each comparison used three independent biological replicates, while the size-distribution analysis included at least 100 microtissues in each group. Results were summarized as mean plus or minus standard deviation. Multiple-group tests used one-way ANOVA followed by Tukey's test, and two-group tests used unpaired two-tailed Student's t-tests, with p < 0.05 treated as statistically significant.
The model was not directly benchmarked against primary human hepatocytes or HepaRG cells, and the simplified microtissues do not reproduce blood flow, oxygen gradients or systemic interactions. Exact effect sizes and confidence intervals were generally not reported. These limits mean the altered markers and CYP activity should not be read as proof that the system matches a human liver.
Research transparency
The research was supported by the Royan Institute through Grant No. 402000234. The authors declared no conflicts of interest related to the work and said the data would be made available on request.
Paper data and sources
Original title: SINEUP-mediated enhancement of HNF4α improves metabolic function in Huh7-based bioengineered hepatic microtissues.
Authors: Zohreh Hashemian, Morteza Zarrabi, Masoumeh Nouri et al.
Journal/Repository: Journal of liver cancer
Status: Peer-reviewed
First online: 2026-08-21
DOI: 10.17998/jlc.2026.05.13
Original paper · Full text