Few previous reports suggested the extra-nuclear intracellular distribution of VDR 
. Among them our work on platelets and megakaryocytes presented the evidences of a mitochondrial VDR 
. In this paper we now demonstrate the mitochondrial localization of VDR also in a human proliferating cell line. Human keratinocytes are not only able to synthesize vit. D but also they depend on vit.D/VDR for proliferation and differentiation 
. We therefore chose to study VDR in the human keratinocytes HaCaT cell line and besides the classical nuclear localization we found an abundant mitochondrial expression of the receptor as well as RXRα its binding partner. VDR distribution was further demonstrated by silencing experiments and was not ligand-dependent: the amount of receptor found in mitochondria was high and not modified by vitamin D. We could define mitochondrial VDR as a resident protein, in contrast with nuclear VDR which is described as a ligand-dependent imported transcription factor. Our work was then focused on identifying the mechanism of VDR mitochondrial import. Since by the most used protein subcellular localization prediction methods we could not find any obvious import sequence at the N-terminal of the receptor, we considered improbable the import through TOM/TIM translocase and we investigated another still underestimated access route to mitochondria: the permeability transition pore (PTP). This channel is formed by a large and dynamic complex of proteins (reviewed in 
), its opening is voltage-dependent and it is more known for being involved in release of mitochondrial proteins such as cytochrome C during apoptosis, in translocation of nucleotides through the complex member adenine nucleotide translocator ANT and in the transport of cholesterol via another complex member, StAR. However very few reports have shown the interaction of mitochondrial proteins with the components of the channel, suggesting the possibility of the import of proteins larger than 1.5 kDa via the PTP. For example the tumor necrosis factor receptor p75NTR was found in mitochondria where interacts with ANT 
. Interestingly, there are reports of a link between PTP opening and mitochondrial localization of proteins. For example mitochondrial localization of p-glycoprotein (pgp) is sensitive to selective opener (Atractyloside glycoside, ATR) and inhibitor (Cyclosporin A, CsA) of mitochondrial permeability transition pore 
. Moreover it has been shown that mitochondrial permeability transition is required for p53 mitochondrial translocation 
In order to verify whether PTP could have a role in VDR import, we first demonstrated by immunoprecipitation studies that VDR and RXRα interacted with PTP by association with VDAC and StAR. Then we searched the proofs that such a multimer was functional. We demonstrated that VDR crossed mitochondrial membranes via PTP by three different sets of experiments: a pharmacological inhibition of PTP opening with CsA, a downregulation of StAR triggered by squalestatin and a decreased expression of StAR upon treatment with dexamethasone. Every time we reduced the PTP function mitochondrial levels of VDR were abated. As for the mitochondrial import of RXRα, the immunoprecipitation studies suggested its association with PTP, however the experiments with CsA indicated that the entry via PTP is a somehow minor event and the main mechanism of import is different, because it is CHX sensitive and only slightly modified by CsA. This is in agreement with previous reports 
. In our experiments mitochondrial p53 was not curtailed by CsA, but on the contrary closing the PTP had an enhancing effect on its mitochondrial localization. These results differ from observations of Liu and coll. 
because they evaluated p53 import triggered by TPA and after short treatments with CsA. In our hands PTP closure for 42 hours stabilized mitochondrial p53. Taken together the results from CHX-CsA experiments were able to discriminate between proteins crossing mitochondrial membranes via PTP or other mechanisms, and they gave the first indication on how VDR enters mitochondria. Moreover the same experiments were important to demonstrate that mitochondrial VDR is a rather stable receptor, showing a half-time of about 48 hours.
VDR entry via PTP was confirmed by the subsequent experiments that decreased the amount of one member of PTP: StAR. In fact both the treatment with squalestatin and with dexamethasone reduced mitochondrial VDR, reinforcing thus the idea that a well functioning PTP was necessary for mitochondrial compartmentalization of VDR. Since the receptor was abundant and stable, a decline in its mitochondrial presence was visible only after a prolonged inhibition of PTP function (48 hours of closure or 72 hours of lower expression). A confirmation of the data obtained by pharmacological down-regulation of StAR came from silencing experiments. StAR knockdown by lentiviral delivery of shRNA decreased mitochondrial VDR, supporting the role of StAR in the mechanism of VDR mitochondrial import.
The presence of nuclear receptors in mitochondria was already described in literature (reviewed in 
) and this localization is important for their role in mitochondrial metabolism. Several papers have described the mitochondrial import and function of other steroid receptors, such as RXRα, estrogen receptor α and β, triiodothyronine (T3) receptor, glucocorticoid receptor 
. For some of them a mitochondrial import signal has been indicated, suggesting a translocase-mediated transport 
sometimes preceded by partial proteolysis 
. However in many cases the import mechanism is still obscure. The presented novel mechanism of entry for VDR could apply to some of them.
In conclusion this paper gives for the first time the evidences of VDR mitochondrial localization in human proliferating cells and shows that mitochondrial targeting is mediated by PTP. These findings open interesting future investigations on PTP function as transporter and on VDR role in mitochondria.