Multiple sclerosis drug FTY-720 toxicity is mediated by the heterotypic fusion of organelles in neuroendocrine cells

FTY-720 (Fingolimod) was one of the first compounds authorized for the treatment of multiple sclerosis. Among its other activities, this sphingosine analogue enhances exocytosis in neuroendocrine chromaffin cells, altering the quantal release of catecholamines. Surprisingly, the size of chromaffin granules is reduced within few minutes of treatment, a process that is paralleled by the homotypic fusion of granules and their heterotypic fusion with mitochondria, as witnessed by dynamic confocal and TIRF microscopy. Electron microscopy studies support these observations, revealing the fusion of several vesicles with individual mitochondria to form large, round mixed organelles. This cross-fusion is SNARE-dependent, being partially prevented by the expression of an inactive form of SNAP-25. Fused mitochondria exhibit an altered redox potential, which dramatically enhances cell death. Therefore, the cross-fusion of intracellular organelles appears to be a new mechanism to be borne in mind when considering the effect of FTY-720 on the survival of neuroendocrine cells.

Confocal microscopy reveals that FTY-720 provokes a reduction in the number of chromaffin granules and concomitant changes in the aspect of the mitochondria. Changes in the vesicular content released following treatment with FTY-720 could be due to altered chromaffin granule size or content and therefore, we studied this possibility using fluorescence confocal microscopy of chromaffin cells expressing the vesicular marker NPY, as indicated elsewhere 29 . Interestingly, after a 10-15 min treatment with FTY-720 (20 µM) there was a reduction in the number of the vesicles labelled with mRFP coupled NPY (red fluorescence, Fig. 2). Measuring the average number of vesicles in cells showed there was a significant reduction after as little as 5 min in the presence of FTY-720, reaching a 30-40% decrease after a 10-15 min treatment. In addition, there was a decrease in the average size of the granules following exposure to FTY-720, which could again be detected as soon as 5 min after incubation with this drug, and that reached a 30% decrease in area (50% reduction in volume) after a 10 min treatment. Consequently, we detected an early change in the formation of microvesicles following incubation with FTY-720 and a concomitant decrease in the number of vesicles 15 min later. These alterations would support the appearance of microvesicle fusions, yet they would appear to contradict the overall increase in the size of the vesicular content released per event. However, it is important to bear in mind the simultaneous change in the mitochondrial structure evident following Mitotracker green labelling ( Fig. 2E), associated with an increase in mitochondrial size (Fig. 2F) and a decrease in their number (Fig. 2G). In addition, these organelles shifted from their characteristic elongated form to a shortened and somewhat rounded appearance after a 15 min treatment with FTY720. This observation was supported by an increase in the mitochondria's roundness (Fig. 2H) and a decrease in their Aspect Ratio (AR, ratio of the longer axis over the shorter one: Fig. 2I). Hence, FTY-720 provokes a clear change in the morphology of the mitochondria that parallels the decrease in the number of chromaffin vesicles. TIRF and confocal microscopy detect homotypic and heterotypic fusion of organelles following FTY-720 treatment. The mitochondrial and vesicle populations are particularly dense in the cortical area of chromaffin cells in culture 30,31 and thus, changes in the destiny of vesicles in the sub-plasmalemma region (up to 150-200 nm from the cell membrane) were tracked by total internal reflection fluorescent microscopy (TIRFM) 32,33 . During the initial 5-10 min after exposure to FTY-720, NPY-mRFP labelled vesicles were apparently smaller in size but later, these fluorescent organelles appeared to become more elongated and larger (Fig. 3A,B). These latter changes were accompanied by a dramatic reduction in the number of vesicles observed, consistent with the confocal microscopy data and with the increase in the size of the labelled organelles measured by mean area determination (Fig. 3E). Consequently, TIRFM experiments revealed an intriguing change in the properties of vesicles, with an initially shift towards the formation of microvesicles followed by their apparent fusion, both homotypic fusion or heterotypic fusion with other major cortical membrane structures, which was Evidence of cross-fusion between chromaffin granules and mitochondria (heterotypic fusion) was obtained by analysing the co-localization of the NPY-mRFP vesicular labelling (red fluorescence in Fig. 3C) and that of mitotracker green in mitochondria (green fluorescence in Fig. 3C) in sequential confocal microscopy images. FTY-720 appears to induce cross-fusion of these organelles as their fluorescent signals appear to co-localize, detected as a yellow colour in the merged images (Fig. 3C). This co-localization was further supported by the Pearson coefficient that correlated the distribution of the red and green fluorescence signals (Fig. 3D), with Electron microscopy confirmed the alterations to chromaffin granules, and the homo and heterotypic fusion of organelles. The co-localization detected above was a strong evidence that FTY-720 induced heterotypic vesicular and mitochondrial fusion, such that we tested these initial observations by performing electron microscopy (EM) on control and FTY-720 treated cells. In electron micrographs of control untreated cells, relatively homogeneous electron-dense chromaffin granules were evident in the cytosol of the cells, and there was a lower density of mitochondria with a weaker electron density, identified by their shape and www.nature.com/scientificreports www.nature.com/scientificreports/ the presence of internal cristae (Fig. 4A). After incubating the cells for 15 min with FTY-720 (20 µM), the size of the electron-dense vesicles was clearly reduced (measuring more than 1,000 vesicles in 11 cells: Fig. 4B), with a 50% fall in their mean circular area (Fig. 4C). In addition, there were 50% fewer vesicles in the cytosol of the treated chromaffin cells (n = 11 cells, P < 0.001: Fig. 4C).
In parallel with the changes detected in the size and number of chromaffin vesicles, we also found that FTY-720 altered the shape and area of the mitochondria (Fig. 4B), which became rounder (Fig. 4D) in conjunction with a four-fold increase in their area (Fig. 4D). Consequently, the EM images supported the data gathered by confocal and TIRF microscopy, proving the simultaneity of the changes observed in chromaffin vesicles and mitochondria when cells were treated with FTY-720.
In the same EM images, some of the events leading to the eventual fusion of organelles were evident. For example, the initial phase of vesicular homotypic fusion was observed (Fig. 4E), with the formation of vesicular contacts between two (Fig. 4F, second row) or three vesicles (third row). Subsequently, there was a redistribution of the dense core of the granules, suggesting a partial loss of the vesicular content that may even reach total release. Following the destiny of these multivesicular fusions, they commonly seemed to reflect complete fusion with mitochondria ( Fig. 4G, second row), involving the fusion of 3 to 5 vesicles per mitochondria and the internalization of the vesicular content in the mixed organelle, without any loss of its compact aspect (Fig. 4G,H, second row). In other images, heterotypic fusion of individual chromaffin vesicles and mitochondria was evident as a possible early event that did not require the prior homotypic fusion of vesicles (Fig. 4H, second row).
After the formation of these mixed organelles, there was evidence of the fusion of these mixed organelles in other images (Fig. 4G, third and fourth row). As such, "giant" round organelles with many characteristics of mitochondria (such as the internal cristae) were ultimately formed, incorporating the dense core of several chromaffin vesicles (Fig. 4G,H, fourth and fifth rows).

The fusogenic properties of FTY-720 are mediated by SNAREs, driving the changes in vesicles and mitochondria. To test whether the alterations to chromaffin cells and mitochondria caused by FTY-720
are mediated by SNARE-dependent fusogenic properties, as demonstrated for exocytosis in different systems 13 , chromaffin cells were transfected with the wild-type SNAP-25 linked to GFP or a truncated SNAP-25 Δ9 that did not prevent SNARE complex formation but reduced secretory activity, mimicking botulinum neurotoxin type A activity 34,35 . In addition, some cells also expressed NPY-mRFP to visualize vesicles, testing the ability of FTY-720 to alter their size and number. Over-expression of the truncated SNAP-25 partially prevented (50% reduction) the formation of microvesicles ( Fig. 5A,C,E) and the reduction in the chromaffin granule population provoked by FTY-720 (Fig. 5D,F). In addition, mitochondrial labelling of the cells (Fig. 5B) demonstrated that SNAP-25 Δ9 over-expression partially prevented the change in area (Fig. 5G,I) and in the roundness of mitochondria (Fig. 5H,J). Together, these results strongly suggest that a late phase of SNARE-mediated fusion is involved in the alterations suffered by these organelles as a result of FTY-720-enhancement of SNARE complex formation 13 .
Mitochondrial alterations affect their redox activity and promote cell toxicity. To assess the functional consequences of the alterations to the organelles produced by FTY-720, we evaluated its effects on the mitochondrial redox potential by measuring the intensity of mitotracker red fluorescence that is sensitive to this parameter 36 . There was an obvious reduction in the Mitotracker red labelling in cells incubated with FTY-720 (20 µM), both in terms of intensity and fluorescence density (Fig. 6A). This reduction in intensity reached around a 70% decrease following a 15 min incubation with FTY-720 (Fig. 6C), indicative of a strong effect on the mitochondrial redox potential.
Impaired mitochondrial function will almost certainly affect cell survival and therefore, we studied the maintenance of cell integrity by Trypan Blue staining (Fig. 6B). As expected, after a 10 min incubation with FTY-720 around 20% of cells took up this dye, indicative of the disruption of their cell membrane (Fig. 6D). Moreover, after a 15 min incubation with FTY-720 (as used in most experiments), followed by a 2 h recovery in fresh medium, we detected 50% cell death relative to the controls. Finally, 24 h after this 15 min incubation even more cell mortality was evident, affecting 80% of the cell population. Thus, it appeared that the rapid and dramatic alterations to mitochondrial morphology caused by FTY-720 via their heterotypic fusion with vesicles was accompanied by a dissipation of their redox potential, which led ultimately to largescale mortality of the cultured cells.

Discussion
FTY-720 a sphingomimetic drug influencing multiple physiological processes in the nervous system. FTY-720, also known as Fingolimod, is a structural analogue of sphingosine with immunosupressant properties 22 that has been approved and is used extensively to treat relapsing remitting MS 12,37 . Its activity is related to the capacity of sphingosine-1P to interact with receptors mediating lymphocyte homing, thereby avoiding immunosupression 38 . Nevertheless, in the last 5 years evidence has accumulated that this compound also influences a variety of processes, including neuronal gene expression, axon growth and degeneration 39 . These effects may not only be associated with the neuroprotective benefits of this drug 40 but also, its therapeutic effects in ischemia, excitotoxicity and memory recovery [41][42][43] . Nevertheless, this drug could also negatively affect neurons, inducing apoptosis by altering calcium signalling and mitochondrial membrane potentials, effects that could be useful to sensitize neuroblastoma cells to anti-.neoplastic drugs 19 .
In terms of the mechanisms that might underlie such a variety of neuronal effects, we recently demonstrated that FTY-720 mimics sphingosine in its activation of synaptobrevin to promote SNARE complex formation, enhancing exocytosis in neuronal and neuroendocrine models at concentrations around 10-20 µM, similar to that found for sphingosine 13 . Here, we show that FTY-720 enhances quantal vesicle release, consistent with the results obtained in rat hippocampal neurons in culture 13 . However, rather than a simple increase in the release of neurotransmitter per event, we observed the formation and fusion of microvesicles, events that ultimately www.nature.com/scientificreports www.nature.com/scientificreports/ provoke an increase in the mean exocytotic release. These complex effects on exocytosis seem to agree well with the effects of sphingosine, enhancing release from microvesicles and full chromaffin granules in bovine 23 and rat chromaffin cells 24 . Interestingly, this effect was also found in lactotrophs 44 , suggesting that the size of vesicles was a major factor influencing the effect of sphingosine, favouring the "kiss and run" fusion of microvesicles and the full collapse of larger dense vesicles.
Our confocal and EM studies clearly show that microvesicles are produced rapidly in the cytosol following FTY-720 treatment, a process that is dependent on SNARE activity. Indeed, some of these vesicles undergo homotypic fusion that could lead to exocytotic fusion, releasing distinct levels of catecholamines, as detected by single cell amperometry.

FTY causes heterotypic fusion of secretory granules and mitochondria.
In searching for possible changes in chromaffin vesicle size to explain the variation in quantal exocytosis, we found important alterations to mitochondrial size and aspect that reflect their heterotypic fusion. Hence, this is the first report of heterotypic fusion linked to the action of an approved drug. The mechanisms underlying this organelle fusion involve SNARE proteins present in the membrane of chromaffin granules. The observation of homotypic granular fusion concomitant with heterotypic fusion during the initial phase of FTY-720 treatment strongly suggests the SNAREs are incorporated into the mitochondrial membranes, in stark contrast to their normal distribution. Subsequently, mitochondria also undergo homotypic fusion to generate giant altered forms of these organelles that include several dense granule cores (normally between 2 and 5). Since this is a late event, it might be related to the incorporation of SNAREs during heterotypic fusion.
A novel mechanism of action of FTY-720 causing mitochondrial impairment and cell death. The data presented here clearly support a new mechanism that explains the effects of FTY-720 on neuronal and neuroendocrine cells, and that is driven by the enhanced fusogenic activity of SNAREs induced by micromolar concentrations of this drug 13 . Accordingly, FTY-720 may diffuse across the plasma membrane www.nature.com/scientificreports www.nature.com/scientificreports/ to accumulate in the cytosol, where it could mimic the activity of sphingosine to promote SNARE complex formation, influencing synaptobrevin availability 13 . The formation of these fusogenic complexes enhances the exocytotic fusion of secretory vesicles, initially augmenting neuroendocrine secretion as occurs with sphingosine derivatives 10,23,24 . However, the concomitant inhibition of F-actin cytoskeleton mediated transport limits the availability of vesicles and their transport from internal pools, as demonstrated here and in accordance with data from astrocytes 28 . Thus, FTY-720 has a complex effect on exocytosis, promoting the release of ready releasable pools of vesicles but impairing the recruitment of other pools from the cell's interior (see scheme in Fig. 7).
Interestingly, SNARE activity is also involved in the formation of microvesicles by membrane fission and homotypic fusion of granules in the cytosol (Fig. 7). The enhanced fusogenic properties of vesicles also leads to the heterotypic fusion of vesicles with mitochondria and finally, to the homotypic fusion of altered mitochondria. As evident in EM images, the alterations to mitochondrial structure are so dramatic as to collapse the www.nature.com/scientificreports www.nature.com/scientificreports/ mitochondrial membrane potential, which probably provokes rapid apoptosis and cell death (Fig. 7). Similar alterations to the redox potential of mitochondria provoked by FTY-720 have been seen in neuroblastoma cells, suggesting this compound induces apoptosis of these tumour cells through various signalling pathways 19 . In conclusion, as far as we know this is the first report of drug induced heterotypic fusion of organelles that induces cell death. This effect must be taken in account when interpreting the effects of FTY-720, compromising the viability of target cells and potentially representing a valuable tool to attack populations of tumour cells of neuronal origin.

Methods
Chromaffin cell preparation and culture. Chromaffin cells were isolated from bovine adrenal glands by collagenase digestion, and they were further separated from the debris and erythrocytes by centrifugation on Percoll gradients, as described elsewhere [45][46][47] . The cells were maintained as monolayer cultures in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% foetal calf serum, 10 μM cytosine arabinoside, 10 μM 5-fluoro-2′-deoxyuridine, 50 IU/ml penicillin and 50 μg/ml streptomycin. The cells were harvested and plated at a density of 150,000 cells/cm 2 in 35 mm Petri dishes (Costar), and they were used between the third and sixth day after plating.
Amperometric determination of exocytosis. To study secretory activity from control non-transfected cells and those expressing the different constructs, the culture medium was replaced by Krebs/HEPES (K/H) basal solution the pH of which was adjusted to 7.4 using NaOH (in mM): NaCl 134, KCl 4.7, KH 2 PO 4 1.2, MgCl 2 1.2, CaCl 2 2.5, Glucose 11, and Hepes 15. Carbon-fibre electrodes insulated with polypropylene and with 14 µm diameter tips were used to monitor catecholamine release from individual chromaffin granules in cells under superfusion 26 . The electrodes were positioned in close apposition to the cell surface using high precision hydraulic micromanipulation, assessing cell membrane deformation with an Axiovert 135 inverted-stage microscope (Zeiss, Oberkochen, Germany) carrying Hoffman optics (Modulation Optics, Greenvale, NY). Electrical connection was accomplished with mercury and an amperometric potential of +650 mV was applied against an Ag/AgCl bath reference electrode using an Axopatch 200 A amplifier (Axon Instruments Inc. Foster City, CA). The current product of catecholamine oxidation was digitized with an A/D converter and recorded at 400 μs/point using the Clampex software (Axon) running on a PC. Experiments were performed in cells stimulated by superfusion with a depolarizing 59 mM high potassium solution (obtained by isosmotically replacing NaCl with KCl), applied through a valve-controlled puffer tip commanded by the acquisition software and located near the cells studied. A curve fitted to the non-linear models provided or implemented in the software (Igor Pro and Graphpad Prism) was used to analyse the data. Individual spike characteristics were studied using the aforementioned Quanta program, allowing for peak detection, integration and kinetic parameter calculations. After acquisition at 2 kHz, only well-defined narrow peaks with an amplitude higher than 5 pA were used to build event histograms. This ensured that the vesicle fusions analysed were produced in the proximity of the electrode and thus, complete www.nature.com/scientificreports www.nature.com/scientificreports/ oxidation of the released catecholamines was accomplished. Cell-to-cell variation was alleviated by using the same electrodes to measure the control cells and those exposed to FTY-720 (20 µM) for 15 min (Sigma Aldrich, SML0700, Lot#056M4724V), as indicated elsewhere 13 . Confocal fluorescence microscopy of chromaffin granules, mitochondria, SNAP-25 and secretion zones. Plasmids were expressed as described elsewhere 48,49  ) to visualize mitochondria, as described previously 30 . MitoTracker labeling was used in transfected (two days post-transfection) and non-transfected cells, and combined with Alexa Fluor 488 (green) to locate active sites of secretion at the plasma membrane by visualizing DBH protein patches. All fluorescent labelling was performed prior to FTY-720 treatment (20 μM: Sigma Aldrich, SML0700, Lot#056M4724V), except the immunolabeling.
For immunolabeling, cells were first incubated with Mitotracker Red (CMXRos), as described above, and they were then treated with FTY-720 for 5 or 15 min prior to studying their basal state and that after stimulation by depolarization with a 59 mM KCl solution for 30 s at room temperature (21-22 °C). Secretion was stopped by lowering the temperature with ice-cold buffer and the sites of secretion were labelled by immunocytochemistry using an anti-DBH rabbit antiserum (1/500: Merck-Millipore, ref AB1585), followed by an Alexa Fluor 488 coupled donkey anti-rabbit IgG (H + L) secondary antibody (1/200: Molecular Probes, ref A21206), again in ice-cold buffer to prevent endocytosis.
Vesicles, mitochondria, SNAP-25 proteins and secretory sites were studied in confocal fluorescence images from equatorial cellular sections, all obtained with an Olympus Fluoview FV300 confocal laser system mounted on an IX-71 inverted microscope incorporating a 100X PLAN-Apo oil-immersion objective with 1.45 NA. Excitation was achieved with argon and helium-neon visible light lasers, and dual fluorescence was registered by sequential acquisition using a 488 nm argon ion (40 mW) to excite EGFP, MitoTracker green and Alexa 488, and with a 543 nm He/Ne 10 mW for RFP and MitoTracker Red CMXRos. A coupled perfusion system was used for FTY-720 treatment, administered at room temperature (22-25 °C) during time-lapse recordings for 15 min.
TIRF microscopy of chromaffin granules. We used transfection of the RFP-NPY plasmid, as described above, to label dense-core vesicles, and two days after transfection live chromaffin cells were recorded by TIRF microscopy after a 5 min incubation with K/H solution followed by 15 min with FTY-720 (20 μM: Sigma Aldrich, SML0700, Lot#056M4724V). All these solutions were applied using the perfusion system at room temperature (22-25 °C).
A through-the-lens TIRFM system was configured on the Olympus IX-71 inverted microscope using a 100× PlanApo 1.45 NA Olympus TIRFM objective 33 . Laser illumination (543 nm He/Ne 10 mW: MellesGriot, Carlsbad, CA, USA) was selected using an Olympus TIRFM IX2-RFAEVA combined system, modifying the angle of laser incidence. In these experiments, laser intensity was kept low to prevent light-induced fusion (2-4% of the maximal intensity) 50 . TIRFM calibration was performed using 100-nm fluorescent beads (Molecular Probes, Invitrogen Detection Technologies, Carlsbad, CA, USA). The depth of penetration for the evanescent field was estimated to be around 200 nm (1/e depth of 180 ± 16 nm), mainly permitting the visualization of the static beads adhered to the coverslip. Fluorescence emission was acquired at a rate of 5 s per frame using an Electron Multiplier charge-coupled device cooled camera (C9100-02 model, Hamamatsu photonics, Japan) and the data was stored on an IBM compatible PC.
Transmission electron microscopy. Bovine chromaffin cell control pellets and pellets pre-treated for 15 min with FTY-720 (20 μM: Sigma Aldrich, SML0700, Lot#056M4724V) were fixed for 2 h at 4 °C with 2.5% glutaraldehyde in 0.2 M cacodylate buffer [pH 7.0]. All the pellets were then washed overnight at 4 °C in a solution of 0.2 M cacodylate buffer, sucrose and distilled water and after post-fixing for 2 h at 4 °C with 1% osmium tetroxide in 0.2 M cacodylate buffer, and extensive washing with distilled water, the samples were stained for 1 h at 4 °C with 2% aqueous uranyl acetate. Subsequently, the samples were washed again, dehydrated through an ethanol series (30,50,70, 80, 96 and 99%: 15 min in each) and incubated twice for 15 min with propylene oxide at room temperature. Finally, the cell pellets were embedded in epoxy resin and ultra-thin sections (70 nm) were obtained on a Leica UC6 ultramicrotome and transferred to copper grids (200 mesh). After staining with uranyl acetate for 5 min and lead citrate for 1 min, the ultrathin sections were analyzed on a JEOL 1011 80 kv transmission electron microscope, with a Gatan BioScam mod 792 digital camera to capture the images.
FTY-720 toxicity on cultured chromaffin cells. We used cell monolayer cultures incubated with supplemented DMEM (see above) and trypan blue solution (1:5: Sigma Aldrich, T6146) cultured for 24 h and evaluated at: t 0 , 5 min, 10 min, 15 min, 1 h, 2 h, 4 h, 8 h, 12 h and 24 h. We prepared control cultures and others that were treated with FTY-270 (20 μM; Sigma Aldrich, SML0700, Lot#056M4724V) for 5, 10 and 15 min. In some cases, after the 15 min FTY-720 treatment the culture medium was replaced by fresh DMEM-Trypan Blue to remove the FTY-720 and the long-term toxicity produced was studied. This evaluation consisted of quantifying cell death (blue cells that incorporated trypan blue through their compromised membrane) and live cells, comparing the control and FTY-720 treated cells at each time points. Cultures were assessed and the cells counted at each time point using an optical microscope coupled to a camera. The numbers of dead and live cells were used to calculate the cell viability (CV, %) for each condition and time, as follows: Statistical analysis. The data are presented as the mean ± SEM (standard error of the mean) and the statistical significance was determined using different tests based on the number of evaluation conditions, the influencing factors and the parametric nature of the data (GraphPad Prism ® 4.0, GraphPad Software Inc., San Diego, CA). A Student's T-test was used to evaluate parametric data with a single influencing factor, and a Mann-Whitney U-Test for two evaluation conditions with a single influencing factor (not assuming a Gaussian distribution). A Kruskal-Wallis Test was used for more than two evaluation conditions with a single influencing factor (not assuming a Gaussian distribution) and a Two way ANOVA Test was used for parametric data where there were two or more evaluation conditions with two influencing factors. The statistical significances are included in the graphs as: n.s, non-significant (P ≥ 0.05); *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001. "N" values reflect the number of cells and organelles analyzed (vesicles and mitochondria) in each condition, and all are also included in the figure legends.
Ethics statement. Adrenal glands were obtained from an industrial slaughterhouse (Matadero de Orihuela SA) that is subject to strict regulations laid down by the Spanish Ministries of Agriculture, Industry and Health, and in accordance with EC guidelines. All the protocols used in this study were approved by the "Organo Evaluador de Proyecto" at the University Miguel Hernández, the office in charge of overseeing the ethical issues associated with animal care and experimentation at our institute.