**3. Discussion**

The presence of mitochondria in the nuclei was claimed more than 50 years ago [1–3], but the objects used for these studies belonged to pathological tissues. This was the reason to assign such a feature to a pathological symptom. In addition, these data were criticized due to the poor quality of the sample and the possibility of artifacts caused by improper treatment of the sample as part of the electron microscopic technique. Penetration of mitochondria into the nucleus as a result of mechanical tissue damage occurring during fixation was discussed by Takemura et al. [10], who found mitochondria in nuclei of myocardial cells taken from patients with various cardiac diseases. However, mitochondria were found in the nucleus of cultured cells, where, due to specific fixation techniques, mechanical damage did not occur.

A reasonable explanation of the presence of mitochondria inside a nucleus was improper execution of mitosis. However, this explanation was not suitable for cardiac myocytes, which belong to postmitotic

cells. Several other observations have recently been made that disprove the assertion that invasion of mitochondria to the nucleus occurs when the nuclear membrane is disassembled during mitosis [23–26].

Confocal microscopy revealed mitochondrial organization in the vicinity of a nucleus, which was di fferent from the well-known interfibrillar and subsarcolemmal mitochondrial population. Mitochondrial web consisting of thin branched filaments covering all peri(intra)nuclear space was typical for all explored normal nuclei of isolated rat ventricular cardiac myocytes. 3D reconstruction of the space occupied by a nucleus demonstrated deep sprouting of mitochondrial filaments into this space (see Supplementary Movie 1). All mitochondrial filaments were fully functional because they were stained with membrane potential sensitive dye and specific mitochondrial marker cardiolipin, and responded to photoexcitation by the partially reversible oscillations of the mitochondrial membrane potential (see Supplementary Movie 2).

Using electron microscopy, we concluded that there was direct contact between mitochondrial clusters and nucleoplasm in cardiomyocytes of the healthy rodents: 3-month-old Wistar rat, 24-month-old OXYS rat, and 5-year-old naked mole rat. Serial ultrathin sections of the same nucleus showed that, depending on the section level, it was possible to observe mitochondria either inside of the closed nucleus or inside of the open nucleus partially devoid of the nuclear membrane. Statistics showed that 1–2% of nuclei present on ultrathin sections of cardiomyocytes contained mitochondria. Our findings are in line with the findings of a Norwegian research group who reported mitochondria in 2–3% of cardiomyocyte nuclei in a patient with rheumatic heart disease [8]. The observations, first made already in the middle of the 20th century of mitochondria inside a nucleus, are no longer considered an artifact of electron microscopy technique [42]. On the basis of a grea<sup>t</sup> number of immunofluorescence assays in which brief disruption of the nuclear membrane in interphase nuclei was observed in association with various diseases and abnormal conditions [18,22–25] as well as in healthy cells [26,43], the presence of mitochondria in the nucleoplasm is usually considered as a result of catastrophic loss of the barrier function of the nuclear membrane that might be a contributing factor of disease progression [44]. In some reports, the penetration of mitochondria into the nucleus was believed to occur due only to a mechanical process [12,44]. It was suggested that the constant contractile function in cardiomyocytes contributes to the penetration of mitochondria into the nucleus through a pathology-weakened nuclear membrane. However, in this study, we showed that mitochondria appeared in the nucleus of normal cardiomyocytes. On the basis of the 3D reconstruction of a part of the cardiomyocyte with nuclei-containing mitochondria, we conclude that the nuclear membrane could be absent in the extensive nuclear region and that it is represented by patches. In all our cases, we describe the presence of mitochondria in nuclei having open configuration without nuclear membrane resealed.

We were unable to answer the question of how specific this discovered phenomenon is for cardiomyocytes. There is evidence that the nuclear membrane undergoes structural changes during mechanical action, which are expressed in local loss of the nuclear envelope integrity [45,46]. This was especially well-observed in the example of migration of cancer cells through narrow holes that led to deformation of the nuclei combined with local breaks of the nuclear membrane [25], which allowed simulating the situation by direct physical actions on the cell [47,48]. Thus, chronic mechanical e ffects on the cell nucleus [49], associated with contractile activity of the heart, could be the cause of similar changes in the structure of the nucleus, leading to local damage/elimination of the nuclear envelope. Cardiomyocytes are cells chronically exposed to a deforming challenge, with mitochondria changing their shape under a normal cardiomyocyte twitch, caused by the dynamic force-balance inside cardiomyocytes and by changes in the spatial sti ffness characteristics [50]. A similar mechanotransduction at the nuclear level was observed in endothelial cells a priori exposed to a chronic shear stress [51].

In 2016, Zhao et al. were the first to study the functional significance of nuclear membrane remodeling in interphase nuclei during erythropoiesis in mice [26]. They showed that this process is not accompanied by a dramatic release of nuclear components into the cytoplasm leading to the loss of cell functions and cell viability, as previously supposed [44]. They showed that the dynamic nature

and cyclic repetition of nuclear opening are essential for normal di fferentiation, ensuring the release of nuclear histones into the cytoplasm for chromatin to be condensed during terminal erythropoiesis. They showed that the release of histones into the cytoplasm is a selective process, and that non-histone nuclear proteins stay inside the nucleus. The released nuclear histones accumulate around the open fragment of the nucleus, performing a protective function, blocking the release of non-histone nuclear proteins, and supporting nuclear/cytoplasmic compartmentalization.

We suppose that local nuclear membrane disassembling, which we observed in cardiomyocytes, as well as subsequent contact between mitochondria and nuclear contents, are of functional significance. Unfortunately, at present, it is impossible to trace the fate of such nuclei and cardiomyocytes containing them. However, during terminal erythropoiesis, Zhao et al. [26] proposed the necessity of the nuclear opening process. As follows from the ultrastructural picture of open nuclei in erythroblasts presented by those authors, the contact between nuclear and cytoplasm components along relatively large areas of the nucleus lacking the nuclear membrane does not lead to cell pathology or apoptosis. In this connection, it is important to mention reports in which authors discovered the direct contact of mitochondria with nuclear components in *Ciona internalis* oocytes [52], as well as with nucleus-like bodies in *Rana pipiens* oocytes [53], and authors have even described special filaments mediating the association of mitochondria with nuclear structures.

It seems that so-called open nuclei, as well as the presence of mitochondria inside nuclei, are a natural and common biological phenomenon related to mitochondrial/nuclear interactions. In eukaryotic cells, mitochondria take part in intracellular regulations mediated by cross-talk between mitochondria and the nucleus. This interaction is represented by anterograde (nucleus–mitochondrion) and retrograde (mitochondrion–nucleus) signaling [32]. This cross-talk includes exchange by ATP/ADP, regulatory proteins and genetic material going in both directions. Bidirectional transport of genetic material is of primary interest for molecular biologists due to its high relevance to the evolution of eukaryotic genomes [27–29] and the occurrence of diseases through regulation of gene expression, possibly by non-coding RNAs originating both from mitochondria [32] and nuclei [31–34,54–57]. The so-called "escape" of nucleic acids from nuclei and mitochondria [58] seems to be part of a well-designed strategy of communication of genomes rather than being an occasional stochastic process. Shortening the distance between genomes will not only facilitate their interaction but also reduce the probability of degradation; in particular, cytosolic nucleases and penetration of mitochondria in the nucleus might serve this strategy.

#### **4. Materials and Methods**

Animals: 3- and 24-month-old male senescence-accelerated OXYS and Wistar rats were obtained from the Shared Center for Genetic Resources of the Institute of Cytology and Genetics (ICG), Siberian Branch of the Russian Academy of Sciences (Novosibirsk, Russia). The OXYS rat strain was established on the basis of the Wistar rat strain at the Institute of Cytology and Genetics, as described earlier [59], and registered in the Rat Genome Database (http://rgd.mcw.edu/). At the age of 4 weeks, the pups were taken from their mothers, housed in groups of five animals per cage (57 × 36 × 20 cm), and kept under standard laboratory conditions (at 22 ± 2 ◦C, 60% relative humidity, and natural light), provided with standard rodent food, PK-120-1, Ltd. (Laboratorsnab, Russia). Naked mole rat colonies were maintained at the Leibniz Institute for Zoo and Wildlife Research, Berlin, Germany, in an artificial burrow system with plexiglass tunnels and boxes. The system was heated to 26–29 ◦C with constant humidity of 60–80%. The chambers contained wood bedding, twigs, and unbleached paper tissue. Fresh food was given daily ad libitum. It included sweet potatoes, carrots, fennel, apples, a cereal supplement containing vitamins and minerals, and oat flakes. The local ethics committee of the "Landesamt für Gesundheit und Soziales", Berlin, Germany (#ZH 156), approved sampling.

#### *4.1. Cardiac Myocytes Isolation*

Ventricular cardiac myocytes used in the study were isolated from adult Wistar rats (2–4 mo old) by a standard enzymatic technique [60] through initial perfusion of hanged isolated heart with the medium containing collagenase type II and subsequent breakage of digested heart pieces by a gentle pipetting and transfer to media with growing Ca2+ content. Final HEPES-bu ffered solution contained (in millimoles per liter) 137 NaCl, 4.9 KCl, 1.2 MgSO4, 1.2 NaH2PO4, 15 glucose, 20 HEPES, and 1.0 CaCl2, pH 7.4.
