Blocking the progression of atherosclerotic lesions and inducing their regression are important clinical goals. We’ve reported several mouse models of atherosclerosis regression in which an improvement of plasma lipoprotein profiles results in a rapid reduction lesion macrophage content. Another common feature was a dramatic phenotypic enrichment of plaque CD68+ cells (macrophages) in markers of the anti-inflammatory (M2) state. Here, we investigate the origin of M2 macrophages in regressing lesions by modifying our original aortic transplantation model to include recipients whose monocytes lack the major receptors responsible for their recruitment into tissues. Aortic arch segments containing an atherosclerotic lesion from an apoE-/- mouse were transplanted into CCR2-/-, CX3CR1-/-, or CCR5-/- mice, in addition to WT and apoE-/- controls. Macrophage marker CD68+ area quantification data showed that the lack of CCR2 or CX3CR1 in transplant recipients completely blocked regression despite the reversal of dyslipidemia, while CCR5 deficiency had no effect. Since inflammation-prone Ly6Chigh monocytes require both CCR2 and CX3CR1 receptors to enter atherosclerotic lesions, our results suggest that the recruitment of new Ly6Chigh inflammatory monocytes is required for regression. Furthermore, CCR2-/- and CX3CR1-/- recipients lacked M2 macrophages in atherosclerotic lesions, suggesting that the M2 macrophages in regressing lesions are derived from the recruitment of inflammation-prone Ly6Chigh monocytes. Finally, we tested whether the M2 enrichment in regression follows the classical Il/4-STAT6 signaling pathway by using STAT6-/- mice as transplant recipients. The lack of STAT6 not only prevented upregulation of M2 markers, but also resulted in impaired lesion regression. In conclusion, we showed that newly recruited Ly6Chigh monocytes give rise to the anti-inflammatory M2 macrophages under regression conditions in a STAT6-dependent manner, and that both the recruitment of new cells and a functioning STAT6 pathway are also required for atherosclerosis regression. The results argue against the possibility that M1 macrophages in progressing lesions repolarize to the M2 phenotype under regression conditions in vivo and support the notion that regression is not a rewinding of progression, but requires the participation of newly recruited inflammation-prone monocytes that in this context promote the resolution of inflammation.