# Novel erythrocyte clumps revealed by an orphan gene Newtic1 in circulating blood and regenerating limbs of the adult newt

## Abstract

The newt, a group of urodele amphibians, has outstanding ability to repeatedly regenerate various body parts, even in the terrestrial life-stage. In this animal, when the limb is amputated, a cell mass named the blastema appears on the stump and eventually gives rise to a new functional limb. Erythrocytes (red blood cells) in most non-mammalian vertebrates, including the newt, preserve their nucleus throughout their life-span, although physiological roles of such nucleated erythrocytes, other than oxygen delivery, are not known. Here we report novel behavior of erythrocytes in the newt. We identified an orphan gene Newtic1, whose transcripts significantly increased in the blastema. Newtic1 was expressed in a subset of erythrocytes that formed a novel clump (EryC). EryC formed a complex with monocytes and was circulating throughout the body. When the limb was amputated, EryCs were newly generated in the stump and accumulated into a distal portion of the growing blastema. Our data suggested that the newt erythrocytes carried multiple secretory molecules including growth factors and matrix metalloproteases, and were capable of delivering these molecules into the blastema as a form of EryCs. This study provides insight into regulations and roles of nucleated erythrocytes, that are independent of oxygen delivery.

## Introduction

The newt belongs to the family Salamandridae in urodele amphibians1. This animal is the only known tetrapod (four-limbed vertebrate) that can regenerate, as an adult in the terrestrial life-stage, their various body parts through reprogramming/dedifferentiation of terminally differentiated cells2,3,4,5,6,7. It is a long-sought issue whether such exceptional ability of the newt can be explained by genes conserved in vertebrates including humans, or by unique genes that the newt may have evolved8,9. In this study, we first investigated unique genes that are specifically expressed in the adult newt limb blastema. We finally screened an orphan gene (we named it Newtic1), that was found only in urodele amphibians. Surprisingly, Newtic1 was specifically expressed in a subset of erythrocytes that formed a novel aggregate structure (we named it the erythrocyte clump (EryC)). We investigated behavior of Newtic1-expressing erythrocytes in normal circulating blood and during limb regeneration in both larval and adult newts. This study provides evidence suggesting that Newtic1-expressing erythrocytes contribute toward limb regeneration specifically in the terrestrial life-stage. Furthermore, to our knowledge, this study provides for the first time insight into regulations and roles of nucleated erythrocytes in non-mammalian vertebrates, that are independent of oxygen delivery.

## Results

### Identification of Newtic1

We first constructed a de novo assembled transcriptome database, named TOTAL, which encompassed genes of the Japanese fire-bellied newt Cynops pyrrhogaster (see Supplementary Figures S1 and S2). This database is now available on the sequence resource site IMORI (http://antler.is.utsunomiya-u.ac.jp/imori/). Utilizing TOTAL and other databases on IMORI, as well as the original mRNA-seq data sets on NCBI (SRP034152), we investigated genes whose transcripts increased in association with blastema formation, and finally identified Newtic1 (1,512 bp; GenBank MG923550) which encodes a 40.7 kD transmembrane protein (Fig. 1; see Supplementary Figures S3 and S4, and Supplementary Table S1). We found a Newtic1 ortholog in another newt species Notophthalmus viridescens, but did not identify any homologous genes in other organisms, including in axolotl (Ambystoma mexicanum) and frogs. Therefore, until recently, this gene was assumed to be specific to the newt. However, recent publications of the genome and transcriptome data sets in urodeles enabled us to discover Newtic1 orthologs in axolotl10 and in another newt species Pleurodeles waltl11 (Supplementary Table S2 and Supplementary Figure S5). Consequently, we conclude that Newtic1 is an orphan gene that might have appeared in a clade of urodele amphibians. In other words, newt-specific protein-coding genes involved in blastema formation are very limited, if present.

### Newtic1 is specifically expressed in a subset of erythrocytes

To identify cells expressing Newtic1, we carried out immunohistochemistry in blastema. Unlike our initial expectation that such cells would be dedifferentiated cells, we wondered if these cells might occur in blood. Therefore, the study was reset by characterizing the blood components (Supplementary Figure S7), and finally concluded that Newtic1 is specifically expressed along the equatorial plane at the periphery of premature erythrocytes (polychromatic normoblasts: PcNobs; Fig. 2).

It is known that vertebrates in good health, except for mammals including humans, have nucleated erythrocytes in blood that circulate throughout the body12,13. Such nucleated erythrocytes correspond to PcNobs or orthochromatic normoblasts (OcNobs) in humans. In adult newts, normoblasts are produced in the spleen and mature in blood vessels during circulation12,13,14,15,16. In peripheral blood, we found various developmental stages of normoblasts (Fig. 2a and Supplementary Figure S7), which accounted for 83–94% in all blood cells (n = 9). The proportion of basophilic normoblasts (immature state) and PcNobs in all normoblasts was less than 2.2% (1.1 ± 0.3%, n = 9) and more than 97.8% (98.9 ± 0.3%, n = 9), respectively. Immunolabeling of blood cells revealed the expression of Newtic1 exclusively in PcNobs, but not all PcNobs were positive (25.4 ± 12.4%, n = 9). Moreover, the proportion of Newtic1-positive (Newtic1+) PcNobs tended to decrease as the PcNobs matured (early stage, 32.0 ± 6.0%; intermediate/late stage, 23.0 ± 4.0%; Student’s t-test, p = 0.0321).

### Newtic1-expressing erythrocytes form an aggregate structure EryC, which is accompanied by monocytes and circulating throughout the body

In tissue sections of normal adult newts, we found that the blood stream in thick vessels contained an aggregate structure of Newtic1+ PcNobs, and these were named the erythrocyte clump (EryC) (Fig. 3). EryCs were typically accompanied by one or two spherical cells that exhibited immunoreactivity to vimentin intermediate filaments at high levels (Fig. 3b–d). The vimentin-immunoreactive (Vim+) cells were identified as monocytes (Fig. 3f–i). It is noteworthy that such EryC-monocyte complexes were frequently observed in the veins of the spleen (Fig. 3a,b), suggesting that normoblasts produced in the spleen remain temporarily in veins to form EryC-monocyte complexes, and the complexes circulate thereafter together with other blood components, including free Newtic1-negative PcNobs. The minimum size of an EryC-monocyte complex is comprised of 1–2 monocytes and 6–10 PcNobs.

### EryCs are newly generated in the limb stump and accumulated in a distal portion of the blastema

We analyzed the changes in the distribution pattern of Newtic1 immunoreactivity during adult newt limb regeneration (Fig. 4; Supplementary Figure S8). In intact forearms of the forelimbs, EryC-monocyte complexes were occasionally observed in thick blood vessels such as radial and ulnar arteries and those along the sub-epidermal layer in the skin, but not in thin blood capillaries within the tissues. No other cells showed Newtic1 immunoreactivity. When the forelimb was amputated in the middle of the forearm, bleeding stopped within a few minutes by coagulation. After bleeding stopped, we carefully removed the clot, which was still soft and jelly-like, from the surface of the wound. At this stage, EryCs were hardly observed in the amputated forearm (day 0 in Fig. 4a). Between 3 and 5 days post amputation (dpa), the wound was closed by the epithelium (wound epidermis) which had grown from the wound edge of the skin of the stump. Following wound closure, a large number of monocytes gathered from blood vessels into the space underlying the wound epidermis, contributing to early blastema (stage I; 5 dpa in Fig. 4a–c). Vim+ cells with long processes were localized on the stump ends of tissues (Fig. 4c). These cells are probably macrophages that are differentiating from monocytes. At this stage, we first recognized a small number of EryCs in the dorsal side of the stump region (5 dpa in Fig. 4a). In this study, we defined the stump region as the region 0–1 mm proximal to the amputation plane. It is important to note that these EryCs were not accompanied by monocytes.

When the top of the blastema started to protrude (stage II), the number of Vim+ cells in the blastema decreased dramatically (14 dpa in Fig. 4a,f). Instead, we first recognized EryCs within the blastema while observing an increase in the number of EryCs in the stump region. As the blastema grew, EryCs in it increased in number and size, and accumulated in the protrusion of the blastema (stage III; 27 dpa in Fig. 4a,d,f). On the other hand, the number of EryCs in the stump region decreased. In this stage, we occasionally observed Vim+ monocytes in the blastema (Fig. 4e). Blastema samples at stages I to III were used to screen Newtic1 (Fig. 1).

Since EryCs from stage I to stage II were not accompanied by monocytes, we attempted to assess their origin. In stage I, we detected Newtic1 immunoreactivity at low levels in a large number of PcNobs that accumulated in dilated blood capillaries in the stump region (Fig. 4g). On the other hand, EryC-monocyte complexes were occasionally observed only in the thick blood vessels near the elbow. In addition, the proportion of Newtic1+ PcNobs in peripheral blood did not change (13–34%, 25.5 ± 4.5%, n = 4). These observations suggest that EryCs are formed on site without monocytes. In other words, EryCs in the growing blastema formed independently of those in circulating EryC-monocyte complexes. In stage III, we observed a large number of Newtic1+ PcNobs that were stuck throughout thick blood vessels in the stump region (Fig. 4h). Monocytes were occasionally recognized in these aggregates. This phenomenon may indicate that the diameter of blood capillaries in the stump region recovered after inflammation ended. This would allow EryCs in the capillaries to be extruded into thick blood vessels where normal blood, which contains EryC-monocyte complexes, circulates.

We addressed how EryCs were translocated from the stump region to the blastema across the amputation plane. A reasonable explanation was transportation through regenerating blood vessels/capillaries. In fact, the spatiotemporal changes in distribution pattern of EryCs seemed to agree with previous descriptions of angiogenesis in the blastema17. Therefore, we carefully examined blastema tissues and confirmed that EryCs were localized within the blood vessels/capillaries extending from the stump to the blastema (Supplementary Figure S8). Interestingly, as the protrusion appeared in the blastema, leading ends of regenerating blood vessels/capillaries in the protrusion connected with each other to form loops with large ventricles. At this stage, EryCs that had passed through different vessels/capillaries gathered to form larger aggregates in the ventricles (Supplementary Figure S8a,b). The presence of EryC-monocyte complexes in the ventricles at stage III may be explained by restarted circulation concomitant with the formation of loops among blood vessels/capillaries.

As the digits appeared in the regenerating forearm, aggregates of Newtic1+ PcNobs showed an obvious decrease in size and density (Supplementary Figure S8c). In the region proximal to the amputation plane, the distribution pattern of Newtic1 immunoreactivity recovered, similar to intact limbs. EryCs observed at this stage were always accompanied by monocytes (Supplementary Figure S8d–f). This may reflect that EryCs originating from PcNobs in the stump region had been replaced with normal blood as circulation recovered.

### EryCs contribute to limb blastema specifically in the terrestrial life-stage

We addressed whether erythrocytes in larval newts also express Newtic1 because amphibians generally switch primary hematopoietic organs from the kidney to the spleen during metamorphosis12,13. In swimming larvae a few weeks before metamorphosis (stage 57–58), we detected Newtic1 immunoreactivity in a small number of erythrocytes in thick blood vessels. These cells rarely gathered and were not accompanied by Vim+ cells (n = 9; Fig. 5). In the limb blastema, large blood vessels containing erythrocytes were observed but Newtic1 immunoreactivity was never detected (n = 3; Fig. 5). These results suggest that larval erythrocytes are capable of expressing Newtic1 and forming aggregates like EryCs, although Newtic1+ erythrocytes do not contribute to the limb blastema. In other words, at least in this species, the contribution of Newtic1+ erythrocytes to the limb blastema is restricted to a post-metamorphic life.

### Erythrocytes carry multiple secretory molecules and are capable of delivering these molecules into the growing blastema as a form of EryCs

We investigated the potential roles of EryCs in blastema formation of the adult limb. It is known that adult newt PcNobs have granulated cytoplasm containing Golgi apparatus and polysomes as well as mitochondria15,16. In addition, our current study on Newtic1 suggests active gene expression in PcNobs. Therefore, we examined the gene expression profile of normal blood by mapping mRNA-seq data (SRP034152; SAMN08574272–08574274) on the TOTAL transcriptome database. The results suggest the expression of a large number of genes encoding secretory molecules (Supplementary Table S3). The list contained growth factors (TGFβ1, IGF-II, BMP2, PDGF-C, VEGF-C, nsCCN) and matrix metalloproteases (Col-a, Col-b, MMP3/10, MMP9, MMP21) (Fig. 6a). We confirmed the transcription of these genes in PcNobs by PCR (Fig. 6b; Supplementary Figure S9), and validated the presence of two growth factors of the TGFβ superfamily (TGFβ1 and BMP2) in PcNobs by immunocytochemistry (Fig. 7). In both growth factors, we found that the level of expression decreased significantly as PcNobs developed (Fig. 7d,e). However, in TGFβ1, expression was higher in Newtic1+ PcNobs throughout development (Fig. 7d), while in BMP2, it was slightly lower in Newtic1+ PcNobs at the intermediate/late stage (Fig. 7e). These results suggest that circulating EryCs, which are composed of Newtic1+ PcNobs at various developmental stages, carry larger doses of TGFβ1 as well as almost average doses of BMP2, relative to the doses carried by the same number of free Newtic1-negative PcNobs. We note that in TGFβ1, the intensity of immunofluorescence was sometimes higher along the equator of PcNobs independent of whether the PcNobs expressed Newtic1 or not.

We carried out immunohistochemistry of regenerating limbs (Fig. 8; Supplementary Figure S10). We found that immunoreactivities to TGFβ1 and BMP2 in the cytoplasm of PcNobs declined as the PcNobs were translocated into the growing blastema as a form of EryCs, while Newtic1 immunoreactivity in those EryC-forming PcNobs was elevated. However, interestingly, in the case of TGFβ1, we frequently observed intense immunoreactivity in a large number of dots distributed along the equatorial plane of PcNobs (Fig. 8a). In the case of BMP2, immunoreactive dots were also observed in areas surrounding PcNobs, but these were mostly localized along the wall of regenerating capillaries/vessels (Fig. 8c). These changes in the immunoreactivity of EryCs and their surroundings were observed only from stage II. It is generally known that TGFβ superfamily molecules are secreted as a latent factor protected by a large protein cage, which binds to extracellular matrices and elastic microfibrils, and contribute to local concentrations of the latent factor18. Taken together, EryCs appear to have released latent TGFβ1 and BMP2 to the internal space of regenerating capillaries/vessels as they entered the region of the growing blastema (Supplementary Figure S11).

## Discussion

In general, circulating erythrocytes are believed to be specialized in the transport of oxygen throughout the body. In adult humans, these cells have lost their nucleus together with other organelles during maturation in the bone marrow, thereby ensuring deformability to pass through even fine blood capillaries in tissues. Nucleated erythrocytes (i.e., normoblasts) are not released in circulating blood in physiologically normal conditions13. On the other hand, in non-mammalian species the nucleus of circulating erythrocytes is preserved12,13. However, there is little information regarding the physiological roles of such nucleated erythrocytes, except for oxygen transportation. Our current study on Newtic1 revealed, to our knowledge, for the first time complex behavior of nucleated erythrocytes in normal blood and during limb regeneration. Our results suggest that PcNobs, a dominant nucleated erythrocyte in circulating blood, carry various secretory molecules throughout the body.

In this study, we found that in the newt, PcNobs are capable of expressing Newtic1 throughout all developmental stages, and EryCs are formed by these Newtic1-expressing PcNobs. However, it remains to be determined how Newtic1 expression is selectively induced in a subset of PcNobs. We speculate that monocytes might contribute to this process because EryCs in normal blood are always accompanied by monocytes, and because, during limb regeneration, Newtic1 expression in the stump region begins concomitantly with the accumulation of monocytes in the space between the amputation plane and the wound epidermis. We have yet to determine whether Newtic1 mediates the aggregation of PcNobs. We also do not yet know the evolutionary origin or functions of Newtic1. Newtic1 is predicted to be a transmembrane protein that is localized along the equator of PcNob. However, since this molecule is unlikely to have an extracellular domain, there is the possibility that Newtic1 may associate with a protein complex that mediates cell aggregation, or that Newtic1 expression may be induced by the same signal as that which exists for cell aggregation. In the latter case, Newtic1 might have different functions from cell aggregation.

The physiological roles of Newtic1+ PcNobs or EryCs also remain to be studied. The dose of secretory molecules that EryCs carry, and when and where EryCs deliver those molecules appear to be rigorously controlled. In fact, EryCs carry larger doses of TGFβ1 and average doses of BMP2 in circulating blood, and seem to deliver both factors within the limb blastema. Interestingly, most of the EryC-carrying secretory molecules (TGFβ1, IGF-II, BMP2, PDGF-C, VEGF-C and matrix metalloproteases) predicted in this study are commonly involved in angiogenesis as well as vascular development and remodeling18,19,20. During limb regeneration, EryCs appear in dilated capillaries in the stump region and change their distribution pattern concomitantly with angiogenesis in the growing blastema. Thus, it is reasonable to propose a hypothesis in which EryCs may work toward the regeneration of blood vessels/capillaries. However, importantly, EryCs do not contribute to the larval limb blastema. Therefore, it is necessary to investigate what functions that are specific to limb regeneration in the terrestrial life-stage require EryCs. In addition, the roles of the EryC-monocyte complex that are also specific to the terrestrial life-stage need to be investigated.

Amphibians are generally able to regenerate their limbs in the aquatic life-stage, however their abilities of limb regeneration consistently decline after they have adapted to the terrestrial environment through metamorphosis. Exceptionally, the newt preserves its high regenerative capacity even in its terrestrial life2,6. Recently, we demonstrated that the newt switches the cellular mechanism underlying limb muscle regeneration from stem cell-based mechanism to dedifferentiation-based mechanism as it grows beyond metamorphosis6. This provided us insight into the presence of mechanisms specific to the terrestrial life-stage. As mentioned above, we demonstrated in the current study that the contribution of EryCs to limb blastema was specific to the terrestrial life-stage, though the presence of newt-specific genes was not evidenced.

In the current study, we listed a lot of candidate secretory molecules that EryCs may deliver (Supplementary Table S3), however we could examine only the two growth factors (TGFβ1 and BMP2) because of availability of antibodies. A recent in vitro study suggests a possibility that BMPs digested by serum proteases such as thrombin and plasmin may promote cell cycle re-entry of dedifferentiating skeletal muscle fibers of adult newt limbs21. In addition, in adult newt limb regeneration, it has been demonstrated that matrix metalloproteases such as Col-a, MMP3/10 and MMP9 are necessary for blastema formation22. As yet, however, it remains to be determined what types of cells provide these secretory molecules. nsCCN, which is a new member of the CCN family growth factor found in the newt, has been assumed to be involved in adult newt heart regeneration23. In this case, endocardial cells in the damaged region express nsCCN. Interestingly, our current results suggest that the expression levels of nsCCN in PcNobs were about ten times higher than those of other cytokines (Fig. 6a). PcNobs or EryCs would be the most likely candidates to provide these molecules to an injured region or to the emerging blastema of a limb. It is important to note that molecules released from EryCs could influence not only blood capillaries/vessels but also mesenchymal cells in the blastema because the permeability of growing blood capillaries/vessels can be regulated24. In fact, we observed neutrophils among mesenchymal cells in the blastema (Fig. 8c). In addition, it is not surprised that when the limb is amputated, tissues on the wound should be exposed to the contents of PcNobs, because PcNobs are released from the wound and are damaged/punctured during coagulation. In the future, functional studies of PcNobs, Newtic1 and EryCs, as well as the candidate secretory molecules, are necessary to further understand the mechanism that enables the newt to efficiently regenerate the limbs in the terrestrial life-stage.

## Methods

All methods were carried out in accordance with Regulations on the Handling of Animal Experiments in University of Tsukuba (RHAEUT). All experimental protocols were approved by Committee for Animal Experiments in University of Tsukuba (CAEUT). Molecular cloning procedures were approved by the University of Tsukuba Safety Committee for Recombinant DNA Experiments (UTSCRDE; 120119, 170110).

### Animals

The Japanese fire bellied newt C. pyrrhogaster was used in this study. Embryonic, larval and adult (total body length: male, ~9 cm; female, 11–12 cm) Toride-Imori25 were used for the construction of transcriptome databases, molecular cloning, and examination of larval blood and limbs. Adult newts that were captured from Okayama, Kyoto, Fukushima and Miyagi Prefecture by a supplier (Aqua Grace, Yokohama, Japan) were also used for the examination of adult blood and limbs. Animals were reared at 18 °C under natural light conditions25. Developmental stages were determined according to established criteria25.

### Anesthesia

FA100 (4-allyl-2-methoxyphenol; LF28C054; DS Pharma Animal Health, Osaka, Japan) dissolved in water was used at room temperature (RT: 22 °C). Before surgery, larval (St. 57–58; ~3 months old) and adult newts were anesthetized in 0.05% FA100 for 15 min and 0.1% FA100 for 1–2 h, respectively6.

### Surgical operations

After anesthesia, animals were rinsed in distilled water and lightly dried on paper towels. To construct transcriptome databases or for histological analysis, animals were sacrificed to obtain tissue samples26; to collect peripheral blood or for the study of limb regeneration, amputation was carried out in the middle of the forearm (at mid-zeugopod region of the forelimb)6. Limb amputation involved amputation of one side of the forelimbs of each animal under a dissecting microscope (M165 FC; Leica Microsystems, Wetzlar, Germany) by a blade (for larvae, a tip of the blade (Cat#: 4991482; Feather Safety Razor, Osaka, Japan); for adults, a surgical blade (No. 14; Futaba, Tokyo, Japan)). Larval and adult amputees were allowed to recover in water and moist containers respectively, and then reared in the same conditions. When the regenerating limb reached a desired morphological stage, we carried out the second amputation at the region of the first amputation or in the middle of the upper arm (at the mid-stylopod region) to obtain regenerates for molecular or histological analysis.

### Construction of de novo assembled transcriptome databases

For this study, a total of 22 databases were constructed from 19 different tissues, and then integrated into one comprehensive database named TOTAL (Supplementary Figures S1 and S2). For each sample, total RNA was purified and qualified (RIN >8.0) to construct a cDNA library26. The cDNA fragments were sequenced (101 bp read x 2) by Illumina HiSeq. 2000/2500 (Hokkaido System Science Inc., Hokkaido, Japan; Sequencing team of KAKENHI 221S0002 in Tokyo University, Kashiwa, Japan)26. Cleaned sequence reads were assembled into contigs (IS-transcripts) by Trinity (version 2013-11-10; https://github.com/trinityrnaseq/trinityrnaseq/wiki; Trinity.pl–seqType fq–JM 196 G–left $read_file_1.fastq–right$read_file_2.fastq–CPU 30–min_kmer_cov 2–output \$output_dirctory_name)26. All information is open to the public through IMORI (http://antler.is.utsunomiya-u.ac.jp/imori/).

### In silico screening of Newtic1

Newtic1 IS-transcripts were obtained through the work flow described in Supplementary Figure S3.

### PCR and molecular cloning

Total RNA was purified from limb tissues or blood samples using Nucleospin kit II (740955.50; Macherey-Nagel GmbH & Co. KG, Düren, Germany), and cDNA was synthesized using Superscript II reverse transcriptase (18064-014; Invitrogen in Thermo Fisher Scientific, Tokyo, Japan) with oligo(dt) 12–18 primers (18418012, Invitrogen in Thermo Fisher Scientific). In the case of limb tissues (Fig. 1g), the mass of blastema samples was weighed and an equivalent mass of intact limb samples was harvested (see Fig. 1a), and these samples were processed for RNA purification and cDNA synthesis at the same scale. Genomic DNA was purified from blood samples using the Wizard Genomic DNA Purification Kit (A1120; Promega, Madison, WI, USA). Using these DNA samples, PCR was carried out with the KODFX system (KFX-101, Toyobo, Osaka, Japan) on an MJ Mini Gradient Thermal Cycler (PTC-1148; Bio-Rad, Hercules, CA, USA). Primer sets used in this study and cycle numbers for the data in figures are listed in Supplementary Table S4. PCR products were subcloned into Escherichia coli using a TA cloning system (45-0640, TOPO TA Cloning Kit, Dual Promoter; Thermo Fisher Scientific) and then sequenced by Sanger protocols (ABI 3130; Applied Biosystems, in Thermo Fisher Scientific)26.

### Preparation of tissue sections and blood cells

For immunolabeling, larval and adult tissues were fixed in modified Zamboni’s fixative (2% paraformaldehyde (PFA)/0.2% picric acid in phosphate-buffered saline (PBS; pH 7.5)) at 4 °C for 6 h27. In some experiments (Supplementary Figure S10), tissues were fixed in 4% PFA in PBS (pH 7.5) at RT for 2 h. Fixed tissues were washed thoroughly with PBS at 4 °C (15 min × 3, 30 min × 3, 1 h, and 6 h), and then allowed to equilibrate in 30% sucrose in PBS at 4 °C. The tissues were embedded in Tissue-Tek® O.C.T. Compound (4583; Sakura® Finetek USA, Inc., Torrance, CA, USA), frozen at about −30 °C in a cryotome (CM1860; Leica), sectioned at about 20 μm thickness, attached on gelatin-coated cover slips, and then dried at RT for a few hours. The sections were stored at −20 °C until use.

Peripheral blood was collected from forelimbs immediately after amputation. For the Wright-Giemsa stain (Fig. 2a and Supplementary Figure S7) and the α-naphthyl acetate esterase stain (Fig. 3i), blood was spread on a glass slide to prepare the blood smear and air dried. For the former stain, the blood on the slide was fixed in approx. 100% methanol for 30 min and air dried. For the latter stain, the blood on the slide was fixed in citrate-acetone-formaldehyde solution (SLBQ6227V; Sigma-Aldrich in Merck, Tokyo, Japan) for 30 sec at RT, and then rinsed thoroughly in running deionized water (DW) for 45–60 sec. These samples were immediately processed for staining (see below). For immunolabeling (Fig. 7a–c), blood was added directly to 4% PFA fixative in a glass bottom dish (D112310; Mastunami, Tokyo, Japan) and then incubated at RT for 2 h. In some experiments (Fig. 2b), blood was fixed in modified Zamboni’s solution at RT for 2 h. Fixed blood cells in suspension were carefully washed with PBS at 4 °C (20 min × 3) and immediately used for labeling. Blood cell suspensions were gently stirred periodically during incubation to prevent coagulation, and solutions were carefully exchanged by micropipettes under a dissecting microscope so as not to lose or select cells. In experiments to examine EryC-monocyte complexes (Fig. 3f,g), blood was put on the bottom of the dish and allowed to coagulate for a few minutes, and then fixed in modified Zamboni’s solution at 4 °C for 6 h. The coagulated blood was carefully washed with PBS at 4 °C (20 min × 3) and immediately used for labeling.

### Blood stain

To characterize blood cells in the adult newt, each blood smear was stained using a standard Wright-Giemsa system (15021; Muto Pure Chemical Co., Ltd, Tokyo, Japan) according to the manufacturer’s instructions. Cell types were identified according to established criteria (Supplementary Figure S6)12,13,16. Monocytes were further characterized using an α-naphthyl acetate esterase staining system (Fig. 3i; SLBQ6227V; Sigma-Aldrich in Merck)28 according to the manufacturer’s instructions. Immunolabeling of blood cell suspensions was carried out as for tissue sections (see below).

### Antibodies

Antibodies used for immunolabeling in this study are listed in Supplementary Table S5.

### Immunolabeling of tissue sections and blood cells

The same procedures for single or double immunofluorescence labeling were applied to both tissue sections and blood cell suspensions27. For single labeling with a certain primary antibody or double labeling with two primary antibodies produced in different species, the following protocol was applied: samples were washed thoroughly (PBS, 0.2% TritonX-100 in PBS, PBS; 15 min each), incubated in blocking solution (5% bovine serum albumin (BSA, 050 M 1599; Sigma-Aldrich in Merck)/2% normal goat serum (S-1000; Vector Laboratories, Burlingame, CA, USA)/0.2% TritonX-100 in PBS) for 2 h, washed as before, and then incubated in primary antibody (antibodies) diluted with blocking solution at 4 °C for 15 h; after washing thoroughly, the samples were incubated in secondary antibody (antibodies) diluted with blocking solution for 4 h and washed thoroughly.

For double labeling with two primary antibodies produced in the same species, entire labeling procedures mentioned above were repeated serially. In the case of double labeling of TGFβ1 and Newtic1, or BMP2 and Newtic1 (Figs 7 and 8), samples were labeled with primary antibody to TGFβ or BMP2 and alexa-488-conjugated secondary antibody. Subsequently, the same samples were labeled with Newtic1 primary antibody and rhodamine-conjugated secondary antibody.

In the case of blood cell suspensions, samples at the bottom of the dish were gently stirred, and solutions were carefully exchanged as for sample preparation. In any labeling protocol, after samples were washed, the nuclei of cells were counterstained with Mayer’s haematoxylin solution (131-09665; Wako Pure Chemical Industries, Ltd., Osaka, Japan), DAPI (1:50,000, D1306; Thermo Fisher Scientific) or TO-PRO-3 Iodide (1:50,000, T3605; Thermo Fisher Scientific). Tissue sections were mounted on a glass slide with 90% glycerol in PBS or into VECTASHIELD mounting medium (H-1000; Vector Laboratories). Tissue sections and blood cell suspensions were immediately subjected to microscopic analysis.

### Western blotting

Protein samples were prepared from blood cells or limb tissues. Cell/tissue samples were collected in lysis solution (25 mM Tris, 150 mM NaCl, 1 mM EDTA-2Na, 1% Igepal CA-630, 1% sodium deoxycholate, 0.1% SDS) containing 1% proteinase inhibitor cocktail (P8340, Sigma-Aldrich in Merck), and protein was extracted according to established protocols27,29. In the case of limb tissues (Fig. 1i), in order to control the mass of tissues, we harvested the same weight of blastema samples and intact limb samples (Fig. 1a) in the same volume of lysis solution. After heat-denaturing in 2× sample buffer, immunoblotting was carried out according to established protocols with some modifications27,29. In brief, proteins (20 μl each) were separated on 10% gel (456–1033, Mini-PROTEAN TGX Precast Gel; Bio-Rad) by SDS-PAGE, and transferred to an Immun-Blot® PVDF membrane (1620–174; Bio-Rad). Membranes were washed in TBST (100 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.05% Tween20) for 10 min, incubated in blocking solution (3% BSA in TBST) containing 2% avidin D (Avidin/Biotin Blocking kit, SP-2001; Vector Laboratories) for 1 h, washed in TBST (1 min × 1, 10 min × 1, 20 min × 1), and then incubated with primary antibody diluted in blocking solution containing 2% biotin (Avidin/Biotin Blocking kit) at 4 °C for 15 h. After washing in TBST (1 min × 1, 15 min × 3), membranes were incubated with biotinylated secondary antibody diluted in blocking solution for 90 min at RT. Membranes were washed in TBST, incubated in AB complex (Vectastain ABC Elite kit, PK-6100; Vector Laboratories) in TBST for 90 min, and then washed in TBST. The immunoreaction was visualized using a DAB substrate kit (SK-4100; Vector Laboratories), and then the membranes were washed in chilled distilled water (5 min × 6).

### Gene expression profiling of blood cells

Total RNA was purified from whole blood samples (n = 3) and mRNA-seq was carried out in a manner similar to the construction of transcriptome databases (see above). To investigate the relative expression levels of genes, sequence reads (deposited to NCBI, SRP034152, SAMN08574272-08574274) were aligned along IS-transcript sequences in TOTAL using the trinity script align_and_estimate_abundance.pl (–est_method RSEM;–aln_method bowtie2). Subsequently, from 168,047 mapped IS-transcripts, we manually screened 304 IS-transcripts encoding secretory molecules including growth factors and MMPs (Supplementary Table S3), and then calculated the mean ± SE of expression levels (normalized FPMK) for each secretory molecule gene (Fig. 6a).

Next, to confirm the expression of these genes in PcNobs by PCR (Fig. 6b), we purified PcNobs (>99%) from whole blood cells (Supplementary Figure S9). Whole blood cells collected in PBS were loaded, at 130–200 μl each, onto pluristrainers (15 μm mesh, 43-50015-01; Pluriselect, Germany) placed in 100 ml beakers filled with 40 ml PBS. The strainers containing blood cells were gently shaken in PBS for 15 min and then blood cells in the strainers were mixed using 500 μl micropipettes. This procedure was repeated three times. Subsequently, 3 ml of blood cell suspensions was carefully transferred onto 3 ml of ficoll solution (30% Ficoll PM400 (Type 400; Sigma-Aldrich in Merck) in PBS) in 15 ml tubes. The tubes were then centrifuged at 100 g for 20 min (LC-120; Tomy Seiko Co., Ltd., Tokyo, Japan). PcNobs deposited on the bottom (about 120 μl) were carefully transferred into a microtube. A small quantity (5 μl) of cell suspension was sampled and recovered in PBS to observe its purity under a microscope (Supplementary Figure S9). The remaining pellet of PcNobs was used for cDNA synthesis followed by PCR analysis (see above).

### Image acquisition and analysis

Bands on agarose gels in PCR and protein bands on membranes in western blotting were scanned and quantified by NIH ImageJ software (https://imagej.nih.gov/ij/). Transmitted light and fluorescence images of tissue sections and blood cells were acquired either by a charge-coupled device (CCD) camera system (DP73; cellSens Standard 1.6; Olympus, Tokyo, Japan) attached to a fluorescence microscope (BX50; Olympus) or through a confocal microscope system (LSM510; LSM 5.0 Image Browser software; Carl Zeiss, Jena, Germany). Microscopic images were analyzed by Photoshop CS5 Extended (Adobe Systems, San Jose, CA, USA) as well as with software for the image acquisition systems. In experiments to quantify the concentrations of growth factors in PcNobs (Fig. 7d,e), we selected representative images (for TGFβ1, six images (total number of PcNobs: 40, 53, 71, 84, 89, 120) from six experiments; for BMP2, three images (total number of PcNobs: 88, 123, 135) from three experiments), that were acquired with a 20× object lens by the BX50-DP73 system (for the field of view, see Supplementary Figure S9). For each image, we measured the amplitude of the mean luminance of the cytoplasm of each PcNob against the background using a function in Photoshop, and normalized the values against the maximum in the same image so that data among experiments could be compared. Figures were prepared using Photoshop. Image, brightness, contrast, and sharpness were adjusted according to the journal’s guidelines.

### Cell counting

In blood smear and cell suspensions, the proportion of cells was estimated by counting the number of cells in representative microscopic images (total cell number/view field >50) obtained from at least three independent experiments. The images were acquired with a 20× object lens by the BX50-DP73 system (see above). In tissue sections, representative images were also obtained from at least three independent experiments. Images covering the area to be analyzed were acquired at high resolution through the confocal microscope system, and merged into one image. In experiments to estimate the proportion of Vim+ cells and the number of Newtic1+ clumps in blastema at 5 and 27 dpa (Fig. 4), we examined representative tissue sections from four individuals at each dpa. For these tissues, we sectioned limb samples along the dorso-ventral axis and used the sections around the central axis of the limb. Only cells that showed a clear nucleus were counted. For Vim+ cells, vimentin-immunoreactive cells with a spherical shape, like monocytes (Fig. 3d,g), were counted. For Newtic1+ clumps (i.e., EryCs), clumps composed of at least six Newtic1+ PcNobs were counted.

### Statistics

Data in the text are presented as the mean ± SE. The statistical mean difference was evaluated by a Student’s t-test, except when data were not homoscedastic, in which case Welch’s t-test was applied, by using Ekuseru-Toukei 2008 software (Social Survey Research Information, Tokyo, Japan).

### Data availability

All relevant cleaned mRNA-seq data have been deposited in NCBI (Sequence Read Archive ID, SRP034152; BioProject ID, PRJNA231688; BioSample ID, SRS515156, SRS589430, SRS685575-685579, SRS685581, SRS685583, SRS685585-685595, SAMN08574272-08574274). The lists of contigs that have been annotated are available in IMORI (http://antler.is.utsunomiya-u.ac.jp/imori/).

## References

1. 1.

Weisrock, D. W. et al. A molecular assessment of phylogenetic relationships and lineage accumulation rates within the family Salamandridae (Amphibia, Caudata). Mol. Phylogenet. Evol. 41, 368–83 (2006).

2. 2.

Alvarado, A. S. & Tsonis, P. A. Bridging the regeneration gap: genetic insights from diverse animal models. Nat. Rev. Genet. 7, 873–884 (2006).

3. 3.

Eguchi, G. et al. Regenerative capacity in newts is not altered by repeated regeneration and aging. Nat. Commun. 2, 384 (2011).

4. 4.

Simon, A. & Tanaka, E. M. Limb regeneration. Wiley Interdiscip. Rev. Dev. Biol. 2, 291–300 (2013).

5. 5.

Islam, M. R. et al. The newt reprograms mature RPE cells into a unique multipotent state for retinal regeneration. Sci. Rep. 4, 6043 (2014).

6. 6.

Tanaka, H. V. et al. A developmentally regulated switch from stem cells to dedifferentiation for limb muscle regeneration in newts. Nat. Commun. 7, 11069 (2016).

7. 7.

Casco-Robles, M. M. et al. Turning the fate of reprogramming cells from retinal disorder to regeneration by Pax6 in newts. Sci. Rep. 6, 33761 (2016).

8. 8.

Kumar, A. et al. Molecular basis for the nerve dependence of limb regeneration in an adult vertebrate. Science 318, 772–777 (2007).

9. 9.

Looso, M. et al. A de novo assembly of the newt transcriptome combined with proteomic validation identifies new protein families expressed during tissue regeneration. Genome Biol. 14, R16 (2013).

10. 10.

Nowoshilow, S. et al. The axolotl genome and the evolution of key tissue formation regulators. Nature 554, 50–55 (2018).

11. 11.

Elewa, A. et al. Reading and editing the Pleurodeles waltl genome reveals novel features of tetrapod regeneration. Nat. Commun. 8, 2286 (2017).

12. 12.

Claver, J. A. & Quaglia, A. I. E. Comparative morphology, development, and function of blood cells in nonmammalian vertebrates. J. Exotic Pet Med. 18, 87–97 (2009).

13. 13.

Giger, K. M. & Kalfa, T. A. Phylogenetic and ontogenetic view of erythroblastic islands. BioMed Res. Int. 2015, 873628 (2015).

14. 14.

Jordan, H. E. & Speidel, C. C. The hemocytopoietic effect of splenectomy in the salamander Triturus viridescens. Am. J. Anat. 46, 55–90 (1930).

15. 15.

Tooze, J. & Davies, H. G. Light- and electron-microscope studies on the spleen of the newt Triturus viridescens; the fine structure of erythropoietic cells. J. Cell Sci. 2, 619–640 (1967).

16. 16.

Pfeiffer, C. J., Pyle, H. & Asashima, M. Blood cell morphology and counts in the Japanese newt (Cynops pyrrhogaster). J. Zoo Wildlife Med. 21, 56–64 (1990).

17. 17.

Peadon, A. M. & Singer, M. The blood vessels of the regenerating limb of the adult newt, Triturus. J. Morphol. 118, 79–89 (1966).

18. 18.

Dijke, P. T. & Arthur, H. M. Extracellular control of TGFβ signalling in vascular development and disease. Nat. Rev. Mol. Cell Biol. 8, 857–869 (2007).

19. 19.

Ucuzian, A. A., Gassman, A. A., East, A. T. & Greisler, H. P. Molecular mediator of angiogenesis. J. Burn Care Res. 31, 158–175 (2010).

20. 20.

Bach, L. A. Endothelial cells and the IGF system. J. Mol. Endocrinol. 54, R1–R13 (2015).

21. 21.

Wagner, I. et al. Serum proteases potentiate BMP-induced cell cycle re-entry of dedifferentiating muscle cells during newt limb regeneration. Dev. Cell 40, 608–617 (2017).

22. 22.

Vinarsky, V. et al. Normal newt limb regeneration requires matrix metalloproteinase function. Dev. Biol. 279, 86–98 (2005).

23. 23.

Looso, M. et al. Spiked-in pulsed in vivo labeling identifies a new member of the CCN family in regenerating newt hearts. J. Proteome Res. 11, 4693–4704 (2012).

24. 24.

Bates, D. O. & Harper, S. J. Regulation of vascular permeability by vascular endothelial growth factors. Vascul. Pharmacol. 39, 225–237 (2002).

25. 25.

Casco-Robles, M. M. et al. Expressing exogenous genes in newts by transgenesis. Nat. Protoc. 6, 600–608 (2011).

26. 26.

Nakamura, K. et al. A transcriptome for the study of early processes of retinal regeneration in the adult newt. Cynops pyrrhogaster. PLoS One 9, e109831 (2014).

27. 27.

Susaki, K. & Chiba, C. MEK mediates in vitro neural transdifferentiation of the adult newt retinal pigment epithelium cells: Is FGF2 an induction factor? Pigment Cell Res. 20, 364–379 (2007).

28. 28.

Yun, M. H., Davaapil, H. & Brockes, J. P. Recurrent turnover of senescent cells during regeneration of a complex structure. eLife 4, e05505 (2015).

29. 29.

Yoshikawa, T. et al. MEK-ERK and heparin-susceptible signaling pathways are involved in cell-cycle entry of the wound edge retinal pigment epithelium cells in the adult newt. Pigment Cell Melanoma Res. 25, 66–82 (2012).

## Acknowledgements

We thank Dr. Takashi Kato (Waseda University) for his valuable comments and discussion about our preliminary observations. We thank all members of the Japan Newt Research Community (JNRC) for their valuable comments and discussion, and the citizens of Toride City for helping us with daily maintenance of Toride-Imori in the field at Imori-no-Sato. This work was supported by grants from the Ministry of Education, Culture, Sports, Science and Technology in Japan (221S0002), the Japan Society for the Promotion of Science (24240062) and University of Tsukuba Basic Research Support Program Type B to C. C.

## Author information

Authors

### Contributions

C.C. and F.T. conceived the idea and evaluated data. R.M.C.-R., F.T. and C.C. designed and performed all experiments, and co-wrote the paper. A.W., K.E., K.T., S.O., T.K., T.A., K.N., T.N. and P.A.T. contributed to collection of RNA samples for transcriptome databases, and evaluated data. M.M.C.-R., K.S. and F.M. contributed to immunolabeling experiments and evaluated data. K.Y. and F.M. supplied and maintained larval and adult newts.

### Corresponding authors

Correspondence to Fubito Toyama or Chikafumi Chiba.

## Ethics declarations

### Competing Interests

The authors declare no competing interests.

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

## Rights and permissions

Reprints and Permissions

Casco-Robles, R.M., Watanabe, A., Eto, K. et al. Novel erythrocyte clumps revealed by an orphan gene Newtic1 in circulating blood and regenerating limbs of the adult newt. Sci Rep 8, 7455 (2018). https://doi.org/10.1038/s41598-018-25867-x

• Accepted:

• Published:

• ### Insights into regeneration tool box: An animal model approach

• Abijeet S. Mehta
•  & Amit Singh

Developmental Biology (2019)

• ### Towards comparative analyses of salamander limb regeneration

• Varun B. Dwaraka
•  & S. Randal Voss

Journal of Experimental Zoology Part B: Molecular and Developmental Evolution (2019)