International Journal of Molecular Sciences Article Cpt1c Downregulation Causes Plasma Membrane Remodelling and Anthracycline Resistance in Breast Cancer Helena Muley 1 , Karmele Valencia 2,3,4 , Josefina Casas 5,6 , Bea Moreno 7 , Luis Botella 1 , Fernando Lecanda 2,3,8 , Rut Fadó 1,9, * and Núria Casals 1,10, * 1 Basic Sciences Department, Faculty of Medicine and Health Sciences, Universitat Internacional de Catalunya, 08195 Sant Cugat del Vallès, Spain 2 Program in Solid Tumors, Center for Applied Medical Research (CIMA), University of Navarra, 31008 Pamplona, Spain 3 Centro de Investigación Biomédica en Red de Cáncer (CIBERONC), 28029 Madrid, Spain 4 Department of Biochemistry and Genetics, School of Sciences, University of Navarra, 31008 Pamplona, Spain 5 Research Unit on Bioactive Molecules (RUBAM), Department of Biological Chemistry, Institute for Advanced Chemistry of Catalonia (IQAC-CSIC), Spanish National Research Council (CSIC), 08034 Barcelona, Spain 6 Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBEREHD), Instituto de Salud Carlos III, 28029 Madrid, Spain 7 Molecular Therapeutics Program, Center for Applied Medical Research (CIMA), University of Navarra, 31008 Pamplona, Spain 8 Department of Pathology, Anatomy and Physiology, University of Navarra, 31008 Pamplona, Spain 9 Institut de Neurociències, Universitat Autònoma de Barcelona, Bellaterra, 08193 Cerdanyola del Vallès, Spain 10 Centro de Investigación Biomédica en Red de Fisiopatología de la Obesidad y la Nutrición (CIBEROBN), Instituto de Salud Carlos III, 28029 Madrid, Spain * Correspondence: rfado@uic.es (R.F.); ncasals@uic.es (N.C.); Tel.: +34-935042000 Abstract: Breast cancer (BC) is the most common malignancy in women worldwide. While the main systemic treatment option is anthracycline-containing chemotherapy, chemoresistance continues to be an obstacle to patient survival. Carnitine palmitoyltransferase 1C (CPT1C) has been described as a poor-prognosis marker for several tumour types, as it favours tumour growth and hinders cells from entering senescence. At the molecular level, CPT1C has been associated with lipid Citation: Muley, H.; Valencia, K.; metabolism regulation and important lipidome changes. Since plasma membrane (PM) rigidity Casas, J.; Moreno, B.; Botella, L.; has been associated with reduced drug uptake, we explored whether CPT1C expression could be Lecanda, F.; Fadó, R.; Casals, N. involved in PM remodelling and drug chemoresistance. Liquid chromatography-high resolution Cpt1c Downregulation Causes mass spectrometry (LC-HRMS) lipid analysis of PM-enriched fractions of MDA-MB-231 BC cells Plasma Membrane Remodelling and showed that CPT1C silencing increased PM phospholipid saturation, suggesting a rise in PM rigidity. Anthracycline Resistance in Breast Moreover, CPT1C silencing increased cell survival against doxorubicin (DOX) treatment in different Cancer. Int. J. Mol. Sci. 2023, 24, 946. BC cells due to reduced drug uptake. These findings, further complemented by ROC plotter analysis https://doi.org/10.3390/ijms24020946 correlating lower CPT1C expression with a lower pathological complete response to anthracyclines Academic Editor: Bernard Haendler in patients with more aggressive types of BC, suggest CPT1C as a novel predictive biomarker for BC Received: 13 November 2022 chemotherapy. Revised: 21 December 2022 Accepted: 31 December 2022 Keywords: CPT1C; chemoresistance; drug uptake; breast cancer; doxorubicin; plasma membrane; Published: 4 January 2023 lipid remodelling; fatty acid saturation Copyright: © 2023 by the authors. 1. Introduction Licensee MDPI, Basel, Switzerland. Breast cancer (BC) is the most common malignancy and the second leading cause This article is an open access article of cancer-related deaths in women worldwide [1,2]. Although BC-related survival rates distributed under the terms and have dramatically improved over the past decades, once BC becomes metastatic, it is conditions of the Creative Commons Attribution (CC BY) license (https:// virtually incurable [3]. The presence or absence of certain molecular markers, namely, creativecommons.org/licenses/by/ estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor 2 4.0/). (HER2), is used to classify BC as follows: hormone receptor (HR) positive/HER2 negative Int. J. Mol. Sci. 2023, 24, 946. https://doi.org/10.3390/ijms24020946 https://www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2023, 24, 946 2 of 18 (HR+/HER2-) BC, occurring in around 70% of patients; HER2 positive (HR+/−/HER2+) BC, accounting for 15–20% of patients; and triple-negative (HR−/HER2−) BC (TNBC), occurring in 15% of patients, a frequently recurring BC, and the most aggressive and lethal BC subtype in women. The receptor signature explains the distinct treatments used for different BC subtypes. For ER+ and PR+ BC patients, endocrine agents are the first-line systemic therapy, in some cases accompanied by chemotherapy. Patients with overexpressed HER2 benefit from targeted therapy, including anti-HER2 antibodies, usually combined with chemotherapy. For TNBC, the only systemic therapy with demonstrated efficacy is chemotherapy [4]. Of the BC chemotherapy drugs, anthracyclines are considered a gold standard, with strong evidence of their positive impact on survival [5]. However, patients treated with anthracyclines commonly develop resistance, and once resistance develops, few treatment options remain [6]. The mechanisms that contribute to the cytotoxic effect of anthracyclines are DNA intercalation, inhibition of topoisomerase II, and the generation of free radicals leading to apoptotic cell death [7]. Cancer cells, in aggressive conditions, can adapt their metabolism to support rapid proliferation, continuous growth, and survival, reprogramming their metabolism to obtain the energy and biosynthetic intermediates that are necessary to preserve their integrity [8]. In recent years, lipids have emerged as central players in cancer cell metabolic reprogram- ming. Changes in the composition and fate of lipids support key oncogenic functions such as cellular energetics, oncogenic signalling, and immune system response attenuation and resistance to stress and chemotherapy [9]. Recent findings have suggested that proteins regulating fatty acid (FA) transport from cytosol to mitochondria, such as carnitine palmi- toyltransferase 1 (CPT1) enzymes, may play a role in cancer cell reprogramming [10–17]. The CPT1 protein family includes CPT1A, CPT1B, and CPT1C. Although CPT1C exhibits high sequence similarity to CPT1A and CPT1B, it is, in fact, a pseudoenzyme with a different molecular function from the canonical CPT1s [18]. CPT1C, which is only expressed in neurons and stem cells in healthy individuals [19], is located in the endo- plasmic reticulum [20] and has very low catalytic activity [21]. In 2011, Zaugg et al. [10] identified CPT1C as a frequently expressed gene in human cancers, including BC and lung cancer. Subsequent CPT1C mRNA expression studies in a wide array of tumour types showed that CPT1C is also highly expressed in brain tumours and several sar- comas [11]. CPT1C overexpression is associated with higher in vitro cell survival and in vivo tumour growth, especially in conditions of metabolic stress (hypoxia or glucose deprivation) [10,12,13,22–25] and also with poorer overall survival [24,26]. CPT1C has also been defined as a novel regulator of cancer cell senescence. Silencing of CPT1C in different tumour cell lines triggers cellular senescence and mitochondrial dysfunction and inhibits tumorigenesis in vivo. In contrast, CPT1C gain-of-function has been shown to reverse cell senescence and to enhance mitochondrial function and FA oxidation (FAO) [12,14,17,22,23]. Those processes have been associated with significant lipidome changes and with lipotox- icity, demonstrating that CPT1C silencing dramatically increases the total fat content of tumour cells, and particularly of certain saturated FAs [22]. Given that lipid saturation changes in the plasma membrane (PM) are associated with drug chemoresistance in several tumours [27–29], we explored whether CPT1C expression could be involved in drug sensitivity loss in BC. We demonstrate that CPT1C silencing drives PM lipid remodelling by increasing phospholipid saturation and chain length and promotes drug impermeability and chemore- sistance in BC cells. We also find that CPT1C downregulation is both a poor-prognosis marker and a poor predictive marker for anthracycline treatment of patients with HER2+ BC and TNBC, i.e., the patients who usually receive chemotherapy. Int. J. Mol. Sci. 2023, 24, 946 3 of 18 2. Results 2.1. CPT1C Silencing Increases Phospholipid Saturation and Chain Length in the PM of MDA-MB-231 Cells To explore whether CPT1C-mediated lipidome changes, as described in the litera- ture [22], could affect the composition of PM in BC cells, we selected the MDA-MB-231 cell line as a model of TNBC [30], and inhibited CPT1C using two short-hairpin (sh) RNAs: shCPT1C_1 and shCPT1C_2 (Figure 1A). We then analysed the lipidome of the PM-enriched cellular fraction in both control and CPT1C-inhibited MDA-MB-231 cells. PM-enriched cellular fraction was obtained by sub-cellular fractionation and was confirmed by western blot (Figure 1B). We found that CPT1C silencing slightly modified the relative abundance of the differ- ent lipid species in PM-enriched fractions of MDA-MB-231 cells (Figure S1A). However, more remarkably, we found a notable difference in phospholipid saturation and chain length, especially for phosphatidylcholine (PC) and phosphatidylethanolamine (PE), the most abundant phospholipids in mammalian cell membranes. The PC species contained longer and more saturated FAs in the CPT1C-silenced cells compared to the control cells. Figure 1C,D depict the levels of PC species found in CPT1C-silenced cells compared to control cells (value = 1). In general, PC species with polyunsaturated FAs were diminished, while PC species with saturated and monounsaturated FAs were increased (Figure 1E); similar results were obtained for the PE species (Figure 1F). For CPT1C-silenced MDA- MB-231 cells, we consistently found enhanced saturation of the FAs that form triglycerides associated with the PM (Figure 1G). We also calculated the PC/PE ratio, a marker of PM fluidity, for the MDA-MB-231 cells, consistently finding that CPT1C silencing significantly decreased this ratio (Figure S1B). Similar to their corresponding precursors, saturated and monounsaturated lysoglyc- erophospholipids were increased in the PM-enriched fractions of shCPT1C-MDA-MB-231 cells (Figure S1C). Considering that the PC and PE species are the most abundant lipids in the PM, and that lipid saturation level and chain length are key determinants of PM fluidity and permeability, our results suggest that CPT1C silencing decreases PM fluidity in MDA-MB-231 cells. As for phosphatidylserine (PS), under CPT1C silencing this was significantly de- creased, independently of the extent of FA saturation (Figure 1H). Since PS has a net negative charge at physiological pH, the fact that it is less present in shCPT1C-MDA-MB- 231 cells may be indicative of a less negatively charged PM, which, in turn, may alter the uptake of positively charged drugs, such as doxorubicin (DOX) in the cancer environment. Even though PM cholesterol is closely related to PM rigidity, total cholesterol lev- els remained unchanged in our CPT1C-silenced cells (Figure S1D). Regarding sphin- golipids, in CPT1C-silenced MDA-MB-231 cells compared to control cells, we found no change in ceramide (Cer), dihydroceramide (dhCer), or sphingomyelin (SM) levels, but did find increased hexosylceramide (HexCer) and dihydrosphingomyelin (dhSM) levels (Figure 1I). Both the dhSM and HexCer species are critical lipid constituents of lipid raft mi- crodomains [31,32], and the dhCer/Cer and dhSM/SM ratios are indicative of the saturation level of the sphingosine backbone. CPT1C silencing increased the dhSM/SM ratio (by 50%), while the dhCer/Cer ratio and the relative SM and Cer proportions remained unchanged (Figure S1E). Considering that dhSM forms, via effective hydrogen bond formation, more condensed and ordered membrane domains than SM, and that SM is much more abundant than Cer in the PM [33], an increased dhSM/SM ratio may contribute to increased PM rigidity in CPT1C-silenced MDA-MB-231 cells. To explore whether CPT1C-mediated lipidome changes, as described in the literature [22], could affect the composition of PM in BC cells, we selected the MDA-MB-231 cell line as a model of TNBC [30], and inhibited CPT1C using two short-hairpin (sh) RNAs: shCPT1C_1 and shCPT1C_2 (Figure 1A). We then analysed the lipidome of the PM-en- Int. J. Mol. Sci. 2023, 24, 946 riched cellular fraction in both control and CPT1C-inhibited MDA-MB-231 cells. PM-en-4 of 18 riched cellular fraction was obtained by sub-cellular fractionation and was confirmed by western blot (Figure 1B). Figure Relative Figure1.1.Relative abundance abundance of lipid of lipid molecular molecular species species in a plasma in a plasma membrane membrane (PM) fraction (PM) fraction of of CPT1C-silenced CPT1C-silenced MDA-MB-231 MDA-MB-231 cells cells measured measured by liquid by liquid chromatography-high chromatography-high resolution resolution mass mass spectrometry spectrometry(LC-HRMS). (LC-HRMS).MDA-MB-231 MDA-MB-231cells were cells infected were with infected withshRandom shRandom or or shCPT1C_2-carry- shCPT1C_2-carrying ing lentivirus. lentivirus. (A)(A)CPT1C CPT1Csilencing silencing was was confirmed confirmedby bywestern westernblot. Tubulin blot. waswas Tubulin usedused as a loading as a loading control. Results are shown as mean ± SD from an experiment performed with 4 replicates (1-way control. Results are shown as mean ± SD from an experiment performed with 4 replicates (1-way ANOVA followed by Dunn’s multiple comparison test; **** p < 0.0001). (B) Western blot showing ANOVA followed by Dunn’s multiple comparison test; **** p < 0.0001). (B) Western blot showing PM enrichment in the isolated fractions (PM) compared with total fraction (TF). Na+ /K+ -ATPase, VDAC-1/Porin, calreticulin, and GAPDH were selected as PM, mitochondria, endoplasmic reticulum (ER) and cytosolic protein markers, respectively. PM fractions show clear enrichment in the PM marker, an almost complete removal of mitochondrial and cytosolic markers, and a slight increase in the ER marker. (C–I) Relative abundance of lipid species in PM-enriched fractions of CPT1C- silenced MDA-MB-231 cells relative to control cells (value = 1, dashed line). Values are grouped by fatty acid (FA) chain length and saturation. In (D,F), relative abundance of specific FAs is shown. (C,D,F) (left). Results are shown as mean ± SD for 3 independent experiments (2-way ANOVA followed by Dunnett’s multiple comparison test; * p < 0.05, ** p < 0.01, **** p < 0.0001, ## p < 0.01, #### p < 0.0001), ns = non significative. (E,F) (right), (G,H,I) Results are shown as mean ± SEM (±SD in H) for 3 independent experiments (Wilcoxon signed-rank test; * p < 0.05, **** p < 0.0001). Int. J. Mol. Sci. 2023, 24, 946 5 of 18 2.2. CPT1C Silencing Impairs DOX Uptake in MDA-MB-231 Cells Int. J. Mol. Sci. 2023, 23, x FOR PEER REVIEW 5 of 18 The intense remodelling of PM lipid composition caused by CPT1C silencing in MDA- MB-231 cells points to a possible decrease in PM permeability to external molecules entering the cells through passive diffusion. The entry of chemotherapeutic agents such as DOX, one as DOX, one of the most efficacious treatments for both early- and late-stage BC, is highly of the most efficacious treatments for both early- and late-stage BC, is highly dependent on dependent on PM lipid composition [34]. For this reason, we explored whether CPT1C PM lipid composition [34]. For this reason, we explored whether CPT1C silencing impaired silencing impaired DOX uptake in MDA-MB-231 cells. DOX uptake in MDA-MB-231 cells. Takingadvantage Taking advantageofofthe theinherent inherentfluorescent fluorescent properties properties of DOX of DOX andand using using flowflow cy- cytom- tometry with the pulse-chase method, we directly analysed intracellular DOX etry with the pulse-chase method, we directly analysed intracellular DOX accumulation in accumula- tion in sh-random sh-random and CPT1C-silenced and CPT1C-silenced MDA-MB-231 MDA-MB-231 cells. Wecells. We incubated incubated cells cells with DOXwithforDOX 4 h, for 4 h, then changed the medium to keep the cells in a drug-free medium then changed the medium to keep the cells in a drug-free medium for a further 20 h. As for a further 20 h. As expected, intracellular DOX accumulation increased in the first 4 h, and expected, intracellular DOX accumulation increased in the first 4 h, and then decreased in a then de- creased in a time-dependent time-dependent manner after manner the mediumafterreplacement. the mediumHowever, replacement. CPT1C However, silencing CPT1C with silencing with both sh-CPT1C sequences (shCPT1C_1 and shCPT1C_2) both sh-CPT1C sequences (shCPT1C_1 and shCPT1C_2) significantly reduced DOX accu- significantly re- duced DOX mulation accumulation in in MDA-MB-231 MDA-MB-231 cells compared tocells compared control to control cells (Figure cellsSupplementary 2A and (Figure 2A and Supplementary Figure S2). Figure S2). Figure Figure 2.2. CPT1C CPT1C silencing silencing decreases decreases doxorubicin doxorubicin (DOX) (DOX) cellular and nuclear cellular and accumulation by nuclear accumulation by de- de- creasing creasing uptake. uptake.MDA-MB-231 MDA-MB-231cells cellswere wereinfected infectedwith withshRandom shRandom oror shCPT1C_2-carrying shCPT1C_2-carrying lentivirus. lentivi- rus.Quantification (A) (A) Quantification of intracellular of intracellular DOXDOXat theatindicated the indicated timestimes as detected as detected by flowby cytometry. flow cytometry.Cells Cells treated were were treated with 1with µg/mL1 µg/mL DOXDOX for 4 for 4 h, after h, after whichwhich the medium the medium was changed was changed to remove to remove DOXDOX and and study drug release until 24 h. Results are shown as median ± SD for 2 study drug release until 24 h. Results are shown as median ± SD for 2 independent experimentsindependent experiments performed with 4 biological replicates (n = 8; 2-way ANOVA followed by Bonferroni’s multiple performed with 4 biological replicates (n = 8; 2-way ANOVA followed by Bonferroni’s multiple comparison test; * p < 0.05, ** p < 0.01, *** p < 0.001). ns = non significative. (B) DOX accumulation comparison test; * p < 0.05, ** p < 0.01, *** p < 0.001). ns = non significative. (B) DOX accumula- kinetics (uptake and normalized release) from both flow cytometry experiments. Slopes for both tion werekinetics compared (uptake (n = 8,and normalized ANCOVA release) for slope from both comparison; flow**** uptake cytometry experiments. p < 0.0001 and release pSlopes for = 0.4366). both were compared (n = 8, ANCOVA for slope comparison; uptake **** p < (C) (left). Intracellular localization of DOX visualized by confocal microscopy 2 h 30 min after DOX 0.0001 and release pexposure. = 0.4366).Nuclei (C) (left). were Intracellular labelled with localization of DOX Hoechst (scale visualized bars, 20 µm).by (C)confocal (right). microscopy Quantification 2 h of30fluo- min after DOXintensity rescence exposure. Nuclei (DOX) inwere labelled the nuclei. with Hoechst Results are shown(scale as bars, mean20 ± SEM (C) 2(right). µm). for Quantification independent experi- ments of performed fluorescence with 4 biological intensity (DOX) in replicates the nuclei.(n Results = 245–289 arecells; shownMann–Whitney as mean ± SEM test; **** for p2