The MRS-detected total choline (tCho) signal is a promising non-invasive surrogate marker of chemotherapy response in breast cancer patients ¹. However, the exact molecular mechanisms by which chemotherapeutic drugs affects the tCho signal, which consists of glycerophosphocholine (GPC), phosphocholine (PC), and free choline (Cho), are often unknown. Here we have used the widely used cancer chemotherapeutic drug doxorubicin to treat estrogen-receptor positive MCF-7 and triple-negative MDA-MB-231 breast cancer cells to elucidate its molecular effects on choline phospholipid metabolism. We employed high-resolution (HR) ¹H MRS to detect changes in cellular choline metabolite profiles, and quantitative RT-PCR and immunoblotting to assess corresponding changes in expression levels of choline-metabolizing enzymes.Cells of the two breast cancer cell lines MCF-7 and MDA-MB-231 were treated with 5 μM doxorubicin, or DMSO vehicle as controls, for 24 hours or 48 hours. Metabolites were extracted using dual-phase extraction (methonal:chloroform:water=1:1:1). HR ¹H MRS of the water-soluble extract fractions was performed on a Bruker 500 MR spectrometer. Choline containing metabolites were quantified from MR spectra using MestReNova software. RNA extracted from 5 μM doxorubicin or vehicle control treated breast cancer cells was reverse transcribed, and SYBR Green based quantitative PCR was used to detect changes in mRNA levels of phospholipase D1 (PLD1), phospholipase D2 (PLD2), choline kinase α (Chkα), glycerophosphodiester phosphodiesterase domain containing 5 (GDPD5) and 6 (GDPD6). PLD1 and Chkα changes in protein levels were detected by immunoblotting following treatment with 0.0 (vehicle control), 0.5, 1.5, or 5.0 μM of doxorubicin. Actin served as loading control. To assess the effects of silencing choline metabolic enzymes on cell migration, MCF-7 or MDA-MB-231 cells were transfected with 75 µM siRNA against PLD1, Chkα, GDPD6, or non-target siRNA using lipofectamine 2000, followed by transwell assay with or without 100 nM doxorubicin in the media, using 2% fetal bovine serium (FBS) as attractant.After 24 hours of doxorubicin treatment of MCF-7 and MDA-MB-231 cells, tCho levels did not change. However, the GPC concentration significantly increased while the PC concentration decreased, thereby dramatically decreasing the PC/GPC ratio as compared to vehicle controls (Fig. 1A-B). This trend even furthered after 48h of doxorubicin treatment, leading to an equal or higher concentration of GPC than PC, dropping the PC/GPC ratio to around 1 (Fig. 1A,C). This was consistent in both breast cancer cell lines. Testing GDPD5 ² and GDPD6 ³, two enzymes implicated in GPC cleavage, we detected that only GDPD6 mRNA was downregulated after doxorubicin treatment (Fig. 2A-B). Chkα, which phosphorylates free choline to PC and regulates cellular PC concentration in breast cancer, was decreased after doxorubicin treatment (Table 1). PLD1, which cleaves choline from phosphatidylcholine, was downregulated by doxorubicin, but not PLD2 (Table 1). Doxorubicin treatment of breast cancer cells resulted in a concentration-dependent decrease of PLD1 and Chkα protein levels, which started to occur at sub-lethal concentrations of 1.5 μM doxorubicin and furthered when the doxorubicin concentration was increased to 5 μM (Fig. 2). A low concentration of 100 nM doxorubicin increased cell migration as also reported previously ⁴, which was abolished by concomitant transient GDPD6 silencing, keeping cell migration down to comparable migration as with transient GDPD6 silencing alone (Fig. 3). Transient silencing of PLD1 by itself reduced migration, but was not able to prevent the doxorubicin-induced increase in migration (Fig. 3).This study demonstrates that while the tCho level may not change following anticancer treatment with a chemotherapeutic agent, the individual components within the tCho signal, such as GPC, PC, or Cho may change. An increase in GPC with a concomitant decrease in PC is a change towards a choline metabolite profile that is typical of nonmalignant breast cancer cells ⁵. This metabolic alteration was caused by doxorubicin-induced decreases in PLD1, Chkα, and GDPD6, which were concentration-dependent as shown for PLD1 and Chkα at the protein level. Our study exemplified that it is worthwhile developing MRS methods such as ³¹P MRS of the breast ⁶ that are able to detect GPC and PC individually for detecting the response to chemotherapy. The first line chemotherapeutic agent doxorubicin was found to promote breast cancer cell migration at low concentration ⁴, which we showed can be counteracted by GDPD6 silencing.The choline metabolite concentrations of GPC and PC and the PC/GPC ratio may serve as non-invasive surrogate makers of therapeutic response in breast cancer patients undergoing chemotherapy with doxorubicin. Targeting GDPD6 may be combined with doxorubicin treatment to minimize its adverse effects on cell migration.