Natural product-derived interventions for reversing multidrug resistance in resistant K562 leukemia cells: a systematic review and meta-analysis

Background Multidrug resistance remains a major barrier to effective chemotherapy for leukemia. Natural product-derived interventions have been widely investigated as modulators of drug resistance in adriamycin- or doxorubicin-resistant K562 cell models, but their quantitative reversal effects remain heterogeneous. Methods In vitro studies evaluating natural product-derived interventions in resistant K562 leukemia sublines were synthesized. The primary meta-analytic endpoint was IC50-based reversal fold (RF), analyzed as log (RF) using an inverse-variance random-effects model. Subgroup, sensitivity, leave-one-out, and small-study-effect analyses were performed. Mechanistic outcomes related to transporter modulation, drug accumulation, apoptosis, and signaling pathways were summarized descriptively. Results Twenty-eight in vitro studies were included, and 19 IC50-based study-level effects were entered into the primary meta-analysis. The pooled RF was 6.31 (95% CI: 4.77–8.36), representing an overall chemosensitization signal across the restricted quantitative dataset. Subgroup analysis showed pooled RF values of 5.71 for K562/A02, 4.90 for K562/ADM, and 8.91 for K562/ADR. Heterogeneity was high (I 2 = 97.7%), indicating substantial variability across resistant-cell sublines, intervention types, substrate drugs, and experimental conditions. Sensitivity analyses confirmed that the pooled effect remained positive, although exclusion of high-effect studies reduced the pooled RF to 4.51. Trim-and-fill analysis produced an adjusted pooled RF of 5.48 (95% CI: 3.87–7.76). Across individual studies, reported mechanisms included ABC transporter modulation, increased intracellular anthracycline accumulation, reduced efflux, restoration of apoptosis, and modulation of PI3K/Akt, NF-κB, and Nrf2/ARE signaling. Conclusion Across the heterogeneous studies, the common IC50-based endpoint showed an overall chemosensitization signal; however, this pooled summary does not imply a shared pharmacological action or potency across chemically distinct metabolites. Effect magnitude and mechanisms remained specific to individual metabolites, resistant-cell sublines, substrate drugs, doses, and molecular targets.