IFITM3 in MKs functions to limit dengue infection. MEG-01s were left alone (No pZIP), or transfected with an empty vector control (pZIP + EV) or an engineered vector containing IFITM3 (pZIP-IFITM3 [eg, IFITM3 overexpression]). IFITM3 mRNA (A) and IFITM3 protein (B) were determined by using quantitative RT-PCR (qRT-PCR) and immunoblot, respectively, to confirm IFITM3 overexpression in unstimulated MEG-01s after transfection. (C) pZIP-EV control or pZIP-IFITM3 overexpressing MEG-01s were infected with DENV2 (t = 96 hours; MOI 0.1) in the presence of subneutralizing concentrations of monoclonal antibodies against DENV2. Antiflavivirus antibody 4G2 (200 ng) was mixed with DENV2 (MOI 0.1) and incubated with MEG-01s. DENV2 mRNA was measured by using qRT-PCR. (D-E) MEG-01s were transfected with control shRNA (sh-NC) or an engineered vector containing shRNA against IFITM3 (sh-IFITM3). sh-NC and sh-IFITM3 MEG-01s were treated with IFNα or vehicle control, and IFITM3 protein was measured (n = 4). (F) sh-NC or sh-IFITM3 MEG-01s were infected with DENV2 (t = 96 hours; MOI 0.1) via antibody-dependent enhancement mode of infection with DENV2 as described earlier. DENV2 mRNA was measured by using qRT-PCR (n = 6). (G-H) A naturally occurring IFITM3 loss-of-function homozygous variant in human MKs increased susceptibility to dengue infection. (G) Representative Sanger sequencing of CD34⁺-derived human MKs harboring either the ancestral IFITM3 allele T (WT, left) or the IFITM3 loss-of-function single-nucleotide polymorphism variant rs12252T-C (IFITM3 variant, right). Red rectangles highlight the single nucleotide change in IFITM3 from a thymine (T, ancestral) to a cytosine (C, variant). (H) MKs harboring either ancestral IFITM3 or the loss-of-function IFITM3 variant (n = 4 different subjects per group) were infected with DENV2 (t = 96 hours; MOI 0.1). DENV2 mRNA was measured by using qRT-PCR and normalized to GAPDH.