Liver Cancer Protein AKR1B10 Could Aid Childhood Tumor Diagnosis, Review Proposes

Diagnosing liver cancer in infants and young children is uniquely difficult, and a new hypothesis-generating review argues that a protein already studied extensively in adult liver cancer deserves systematic attention in pediatrics. Writing in Frontiers in Medicine, a team of Chinese and Japanese researchers proposes that aldo-keto reductase 1B10, or AKR1B10, a secreted enzyme repeatedly validated as a blood marker for adult hepatocellular carcinoma, may also prove useful in hepatoblastoma and other childhood liver malignancies. The authors are careful to stress that no study has yet measured AKR1B10 in any pediatric liver tumor; their paper is a theoretical framework built on adult evidence and shared molecular pathways, not a clinical finding.

The clinical problem the review addresses is well documented. Serum alpha-fetoprotein, or AFP, combined with abdominal ultrasound remains the most widely used non-invasive screening protocol for liver cancer in both adults and children. Yet AFP performs poorly for early disease. In adults, its sensitivity for early-stage hepatocellular carcinoma is only 44.4 percent, with a specificity of 84.8 percent and an area under the curve of 0.52, according to studies cited in the review. Abdominal ultrasound detects early-stage disease with a sensitivity of roughly 45 percent, and even when combined with AFP, sensitivity rises only to about 63 percent. Between 30 and 40 percent of all hepatocellular carcinoma cases, and 60 to 70 percent of early-stage cases, are AFP-negative, and cirrhosis can produce false-positive AFP results.

In children, the obstacle is different but equally serious. AFP is produced by the fetal liver and embryonic yolk sac, and after birth its levels decline rapidly, typically reaching adult values within six to twelve months. In children under three, and especially infants under one year, the normal range varies enormously from child to child and shifts with age. Hepatoblastoma, the most common pediatric malignant liver tumor, accounting for roughly 80 percent of such tumors with a peak incidence below three years of age, often drives AFP to peaks of up to 1,000,000 ng/mL. But in small-cell undifferentiated hepatoblastoma, AFP levels may be normal or only mildly elevated, below 100 ng/mL, making those cases easy to miss. Without standardized, age-stratified AFP reference ranges, distinguishing pathological elevation from physiological fluctuation is genuinely hard, and the review notes this can lead to unnecessary biopsies, delayed treatment, or misdiagnosis.

The stakes of early detection are high in both age groups. The five-year survival rate for advanced hepatocellular carcinoma is below 15 percent, while curative resection of early-stage disease yields five-year survival of 69 to 86.2 percent. In children, the three-year overall survival rate for high-risk hepatoblastoma is only 52 to 65 percent, compared with more than 90 percent for standard-risk disease. Against this backdrop, AKR1B10 has accumulated a substantial adult evidence base. The protein consists of 316 amino acids, weighs approximately 36 kDa, sits on chromosome 7q33, and shares 70 percent sequence homology with AKR1B1 and 91.5 percent with AKR1B15. It is an NADPH-dependent reductase whose substrates include retinaldehyde and cytotoxic aldehydes generated by lipid peroxidation, and among AKR1B family members it shows the highest catalytic efficiency for retinal, making it the most potent retinal reductase currently known.

Several features make AKR1B10 attractive as a blood marker. It is a secreted protein, released from normal intestinal epithelial cells and cancer cells through a lysosome-mediated non-classical pathway, with concentrations in ileal fluid reaching 188.6 to 535.7 ng/mL. Its expression in hepatocellular carcinoma follows a distinctive pattern: high in early-stage, well-differentiated tumors and lower in advanced, poorly differentiated ones, which underpins its sensitivity for early detection. A multicenter study by Ye and colleagues found that the area under the curve for serum AKR1B10 in detecting hepatocellular carcinoma exceeded that of AFP, with the advantage most pronounced in early-stage and AFP-negative disease, and combined use of the two markers improved diagnostic accuracy. A meta-analysis cited in the review reported an overall specificity of 78 percent and sensitivity of 85 percent for AKR1B10 in diagnosing hepatocellular carcinoma, with performance for early-stage disease surpassing AFP and serum levels declining rapidly after surgery, allowing treatment response monitoring.

The review also documents AKR1B10’s mechanistic roles in tumor biology. By catalyzing the reduction of retinaldehyde to retinol, it competes with retinoic acid synthesis, depleting a signal that maintains cell differentiation and thereby promoting dedifferentiation and proliferation. It stabilizes acetyl-CoA carboxylase alpha, the rate-limiting enzyme in fatty acid synthesis, although the authors note this mechanism has been demonstrated in breast cancer cells and not yet directly confirmed in hepatocellular carcinoma. In liver cancer cells, AKR1B10 also promotes proliferation through sphingosine-1-phosphate signaling and by activating PI3K and AKT phosphorylation, and it is transcriptionally regulated by beta-catenin signaling. More recently, in lenvatinib-resistant hepatocellular carcinoma, AKR1B10 has been shown to drive resistance by enhancing aerobic glycolysis through a lactylation-mediated positive feedback loop involving lactate dehydrogenase A, a process that the AKR1B10 inhibitor epalrestat can reverse. The enzyme also contributes to resistance against cisplatin and doxorubicin through carbonyl detoxification.

Clinical applications in adults extend beyond diagnosis. Immunohistochemical studies have found AKR1B10 overexpressed in 97 percent of hepatocellular carcinoma tissues but virtually absent in benign lesions such as hepatic adenomas and focal nodular hyperplasia, aiding differential diagnosis. In two reported cases of hepatitis B-related hepatocellular carcinoma with persistently negative AFP, serum AKR1B10 was markedly elevated, supporting its role as an auxiliary marker for AFP-negative disease. In patients with chronic hepatitis C, those with high AKR1B10 expression in the liver before treatment had roughly a 15-fold higher risk of subsequently developing hepatocellular carcinoma even after achieving sustained virologic response, with a negative predictive value of 0.991 that could spare low-risk patients frequent surveillance. Prognostic findings, however, conflict: some studies associate high expression with better survival and lower recurrence, while others link it to poor outcomes, a discrepancy the authors attribute to differences in etiology, stage, and the enzyme’s dual protective and pro-tumor roles.

The bridge to pediatrics rests on shared molecular pathways. Hepatoblastoma is an embryonic tumor with a low somatic mutation burden, averaging about 3.9 mutations per tumor, and activating CTNNB1 mutations are its core driving event, present in roughly 70 to 90 percent of cases, compared with about 27 to 35 percent in adult hepatocellular carcinoma. Because AKR1B10 is regulated by beta-catenin signaling in adult tumors, the authors hypothesize it may also be upregulated in hepatoblastoma, potentially helping to resolve the diagnostic blind spots created by physiological AFP fluctuations in infancy, and may serve as an auxiliary serum marker for AFP-negative or low-expressing hepatoblastoma subtypes. Similar reasoning extends to pediatric hepatocellular carcinoma, which shares Wnt/beta-catenin, TP53, and TERT promoter abnormalities with the adult disease, and to subtypes such as fibrolamellar carcinoma and hepatocellular neoplasm not otherwise specified, where AFP is often not elevated.

The authors are explicit that these proposals remain purely hypothetical. Shared pathway abnormalities do not guarantee comparable AKR1B10 expression, regulation, secretion, or diagnostic utility in children, whose tumors differ in developmental origin and mutation profile. Normal serum ranges for children, especially infants and toddlers, have not been established, and levels may be influenced by age, sex, puberty, inflammation, liver damage, renal function, obesity, and medications. Adult diagnostic thresholds, such as the cutoff of at least 1,078.2 pg/mL used for nonalcoholic steatohepatitis, cannot be directly extrapolated. Transcriptomic work showing that neonates carry 669 overexpressed and 440 underexpressed genes and 36,782 differential chromatin accessibility peaks relative to adults reinforces the point that adult mechanisms cannot simply be assumed to hold in children.

To move the hypothesis forward, the review lays out a four-stage, explicitly exploratory validation framework: first, characterizing AKR1B10 expression in hepatoblastoma tissues and pediatric controls using immunohistochemistry and RT-qPCR, stratified by histological subtype; second, establishing age- and sex-specific serum reference ranges in healthy children and children with benign liver disease; third, case-control assessment of diagnostic performance comparing AKR1B10 alone, AFP alone, and their combination in independent pediatric cohorts; and fourth, exploratory evaluation of associations with treatment response, recurrence, and chemoresistance using hepatoblastoma cell lines and patient-derived xenograft models. The authors position AKR1B10 strictly as an adjunctive marker to be interpreted alongside age-adjusted AFP, imaging, PRETEXT staging, and pathology, not as a standalone diagnostic tool or a screening test for the general pediatric population. They also flag open questions on the pediatric pharmacokinetics and safety of epalrestat and on selective off-target inhibition given the 70.89 percent sequence homology between AKR1B10 and AKR1B1. Until direct pediatric data accumulate, the proposal remains what the authors intend it to be: a testable research direction rather than a validated clinical advance.

Subject of Research: The potential diagnostic value of the AKR1B10 protein in pediatric liver malignancies based on adult hepatocellular carcinoma evidence

Article Title: Diagnostic potential of Aldo-keto reductase 1B10 in pediatric hepatic malignancies: a narrative, hypothesis-generating translational review based on adult evidence and future research directions

Article References: Luo, J., Liao, P., Zeng, Z., Deng, Q., Hu, M., Liu, L., Huang, C., Yin, Q., Peng, J., Cao, D., Hu, X., & Chen, L. (2026). Diagnostic potential of Aldo-keto reductase 1B10 in pediatric hepatic malignancies: a narrative, hypothesis-generating translational review based on adult evidence and future research directions. Frontiers in Medicine, 13, Article 1972682. https://doi.org/10.3389/fmed.2026.1972682

Image Credits: AI Generated

DOI: 10.3389/fmed.2026.1972682

Keywords: AKR1B10, hepatoblastoma, hepatocellular carcinoma, alpha-fetoprotein, pediatric liver cancer, biomarker, Wnt beta-catenin pathway, retinoic acid signaling, early diagnosis, serum marker, chemoresistance, narrative review