KIFC1 is essential for normal spermatogenesis and its depletion results in early germ cell apoptosis in the Kuruma shrimp, Penaeus (Marsupenaeus) japonicus

The main features of P. japonicus kifc1

The full-length kifc1 of P. japonicus is 2650 bp (GenBank accession number: MN072915). It contains 136 bp 5′ untranslated region (UTR), 2217 bp open reading frame (ORF), and 297 bp 3′ UTR. It encodes 738 amino acids and the predicted molecular weight is 81.10 kDa (Figure 1). The isoelectric point of this protein predicted by ExPASy-ProtParam tool is 9.47 (https://web.expasy.org/protparam/).

The secondary structure prediction of P. japonicus KIFC1 showed three primary domains. A divergent tail at the amino terminal is from 1 to 192 amino acids, a coiled stalk begins from 193 to 377 amino acids, and a head at the carboxyl terminal is composed of the 378-738 amino acids, which contains the conserved motor domain and “walks” along the microtubule (Figure 2A). Additionally, we predicted the putative tertiary structure of KIFC1 protein, in which all three domains’ 3-D structures were observed clearly (Figure 2B–2E).

We aligned KIFC1 of P. japonicus with its homologues of other species and found it has 66.1%, 57.7%, 35.5%, 36.1%, 35.8%, 36.6%, 35.5% and 33.1% identity with its homologues in Eriocheir sinensis, Macrobrachiumnipponensis, Danio rerio, Bos taurus, Homo sapiens, Mus musculus, Gallus gaullus, and Alligator sinensis, respectively. The alignment result also showed a more conservative head region in KIFC1 when compared with the stalk and tail (Figure 3). In addition, there are three ATP binding sites (LAGSE, SSRSH and AYGQTGSGKT), a microtubule binding site (YNETIRDLL) and a KIFC conserved domain (ELKGNIRVFCRVRP) in the head region of P. japonicus KIFC1 (Figure 3). The phylogenetic analysis revealed the putative KIFC1 of P. japonicus constitutes a sister clade with it homologues of Procambarus clarkii, and also suggested it has a closest genetic relationship with P. clarkii among the examined species in this study (Figure 4).

KIFC1 is highly expressed in reproductive system of P. japonicas

The expression of kifc1 mRNA in different tissues of P. japonicus was determined using semi-quantitative RT-PCR. A 384 bp fragment of kifc1 cDNA and a 179 bp fragment of β-actin were amplified in the heart, hepatopancreas, muscle, gill, vas deferens, spermatophore and testis (Figure 5A). β-actin was served as an internal control. Gray analysis by the Image J software indicated that kifc1 mRNA was extensive expressed in all selected tissues, and the high expression occurred in the testis, vas deferens and spermatophore (Figure 5B). Meanwhile, we identified and analyzed KIFC1 protein expression in muscle, heart, testis, vas deferens and spermatophore by Western blots. A band about 81 kDa was recognized in all of these samples (Figure 5C). The expression trend of KIFC1 protein in all tissues is almost consistent with that of kifc1 mRNA (Figure 5D), which implies that KIFC1 may have important roles in spermatogenesis of P. japonicus.

KIFC1 participates in mitosis, meiosis as well as spermiogenesis of P. japonicas

In order to explore the role of KIFC1 and microtubules during spermatogenesis of P. japonicus, we conducted the immunofluorescence further to examine the localization of them. In the spermatogonial phase, KIFC1 co-localized with tubulin in cytoplasm. Meanwhile, KIFC1 also dispersed in the nucleus (Figure 6A, arrows). In the spermatocyte phase, the signal of KIFC1 and tubulin in cytoplasm was obviously increased when compared with the spermatogonial phase. More importantly, a large number of KIFC1 signal was detected in the nucleus of spermatocyte (Figure 6B, arrows). In the early stage of spermatids, the microtubules began to assemble to one side of the nucleus (Figure 6C, arrows). The KIFC1 signal was exported from the nucleus, and spread over and co-localized with the microtubules in the cytoplasm (Figure 6C, arrows). In the late-stage spermatids, the microtubules exhibited characteristic spindle morphology around the nucleus (Figure 6D, arrows), while the signal of KIFC1 in this phase presented a relative increased (Figure 6D). In the control group, no signal of KIFC1 and tubulin was detected from the spermatogonia to late spermatids (Supplementary Figure 1). These results suggest that KIFC1 is likely to play essential roles during spermatogenesis of P. japonicus including mitosis of spermatogonia, meiosis of spermatocytes and the acrosome formation and maintaining.

During post-spermiogenesis in vas deferens (Figure 7), the microtubules also showed spindle-like structure around the nucleus (Figure 7B, arrows), while the signal of KIFC1 in this phase presented a relative reduction (Figure 7C). In mature sperm from the spermatophore, we can observe the nucleus, acrosome and spike under phase contrast microscopy (Figure 7I, arrows). The signal of microtubules and KIFC1 was completely not co-localized (Figure 7J, arrows). KIFC1 was localizes mainly in the acrosome, but the microtubules was close to the nucleus at the opposite end of acrosome (Figure 7G, 7H, 7J, arrows). These data show that KIFC1 has almost finished its task during spermatogenesis of P. japonicus, and continues to maintain the acrosome morphology in mature sperm in spermatophore.

KIFC1 knockdown has a severely impact on microtubule distribution and results in early germ cell apoptosis in the testis of P. japonicus

In order to explore the function of KIFC1 during spermatogenesis of P. japonicus, we conducted the RNAi experiment through the injection of dsRNA to knockdown the expression of KIFC1 in vivo. The knockdown efficiency of kifc1 was examined by semi-quantitative RT-PCR and Western blot (Figure 8A–8D). From the results of gray analysis, both dsKIFC1-1 and dsKIFC1-2 group had a significantly lower kifc1 mRNA and KIFC1 protein expression than those of control groups (Figure 8B, 8D). Therefore, both of these two RNAi groups are feasible for further studies.

To determine whether there are effects on microtubule assembly, we observed the distribution of microtubules in the dsKIFC1-1 testis, vas deferens and spermatophore. In the dsKIFC1-1 testis, the signal of KIFC1 is hardly detectable. At the same time, the microtubule signal has an obvious decrease and it presents an abnormal accumulation in spermatogonia (Figure 9A, and Supplementary Figure 2A). However, in the dsKIFC1-1 vas deferens and spermatophore, although KIFC1 expression is knocked down, the expression and distribution of microtubules are notaffected (Figure 9B, 9C, and Supplementary Figure 2B, 2C). These results indicate that KIFC1 knockdown has a serious influence on the distribution of microtubules during the early spermatogenesis of P. japonicus including mitosis and meiosis, but not in the late stage, spermiogenesis.

In order to examine the effect of KIFC1 knockdown on spermatogenesis in P. japonicus, we conducted TUNEL assay. In the control groups (Normal and dsEGFP testis), no apoptosis is detected (Figure 10A, 10B). But the different results occur in the RNAi groups. Plenty apoptosis is detected in dsKIFC1-1 testes (Figure 10C), and apoptosis is taken place in spermatogonia and spermatocytes at high magnification (Figure 10F). In addition, we also detected apoptosis in the RNAi vas deferens and spermatophore, but the result shows that KIFC1 knockdown has no effect on the spermatids in vas deferens and the mature sperm in spermatophore, which is identical with that of microtubule assembly (Figure 10D, 10E).

KIFC1 depletion affects the distribution of microtubules and induces the early germ cell apoptosis during spermatogenesis of P. japonicas

KIFC1 transports various cargos, including membrane-bounded organelles, vesicles and interacting proteins, along the microtubule from the plus end to the minus end [30, 39]. Additionally, since the KIFC1 contains a conserved nuclear localization signal in the tail domain, it can be transferred into the nucleus and participates in the regulation of mitosis and meiosis [40–42].

The mitotic and meiosis spindles are molecular machines composed of a bipolar array of microtubules, which contributes to chromosome segregation via microtubule sliding, growth as well as microtubule shrinkage [42, 43]. KIFC1 participates in spindle assembly and modulates spindle length [44]. In many kinds of cancer cells, KIFC1 has been demonstrated playing an essential role in bipolar spindle formation by clustering the multiple poles and maintaining chromosomal stability [45, 46]. Therefore, KIFC1 is closely associated with the dynamic of microtubules during both mitotic and meiosis. In order to explore whether KIFC1 knockdown has an impact on the microtubule dynamic during spermatogenesis of P. japonicus, we conducted RNA interference by dsRNA injection in vivo. From our results, decreased expression of KIFC1 leads to low microtubule signal and abnormal accumulation just in the early stages of spermatogenesis in P. japonicus testis. We speculated that the main reason for this phenomenon may be that special morphology assembled by microtubules has been accomplished before spermatids were released into the vas deferens (Figure 9).

KIFC1 is essential for mitosis not only in normal cells, but also in cancer cells. The KIFC1 deficiency affects the normal mitosis, as well as damages cell growth, proliferation and survival. Wei and Yang [47] found that the KIFC1 knockout 293T cells are hard to heal from scratched mechanical damage and cell death occurs in the wound healing assay. In KIFC1-depleted IMR-90 cells, although no obvious immediate cell death was observed, 20% less cells were detected compared with control group. Interestingly, when knocked down both KIFC1 and Mad2 in IMR-90 cells, the cell apoptosis was triggered suggesting their combined effect during mitosis [45]. KIFC1 presents upregulated in various cancer cells and is related to cancer progression [31, 48, 49]. Besides, the overexpression of KIFC1 can inhibit docetaxel-mediated apoptosis in breast cancer cells and prostate cancer cells [49, 50]. Several studies have shown that KIFC1 is associated with an apoptosis pathway, and inhibition of KIFC1 suppresses cell proliferation, but also induces the apoptosis pathway in different cancer cells [31, 49, 51]. In our study, we determined that cell apoptosis commonly occurred in the spermatogonia of the KIFC1 knockdown testis, which suggested the key function of KIFC1 in mitosis of spermatogonia and KIFC1 depletion resulting in spermatogonia apoptosis in P. japonicus (Figure 10).

In meiotic germ cells, KIFC1 also has a role in the regulation of spindle morphology and length, and in chromosome alignment and segregation. KIFC1 inhibition in mouse oocytes leads to aberrant spindle assembly and finally generates meiosis arrest [52]. Ncd, a KIFC1 orthologue in Drosophila, is reported to form the microtubule nucleating center at the spindle pole body during meiosis in oogenesis [53, 54]. In addition, KIFC1 depletion in Procambarus clarkii causes a whole suppression of spermiogenesis, which is speculated for the abnormal early spermatids undergo apoptosis due to an unsuccessful meiosis, or as a result of incomplete structure in KIFC1 knockdown testis [55]. All these results suggested that kifc1 ablation resulted in the normal meiosis interruption or finally induced apoptosis in germ cell genesis [54, 55]. In this study, we detected that in the spermatocyte phase, the signal of KIFC1 and tubulin was obviously increased in cytoplasm by immunofluorescence. Besides, abundant KIFC1 molecules were transported into the nucleus, which implied its importance for normal meiosis of male P. japonicus (Figure 6B, arrows). Further analysis of the function of KIFC1 by TUNEL assays showed that KIFC1 knockdown triggered cell apoptosis in spermatocyte phase (Figure 10), which gave the evidence for its important role in normal meiosis during spermatogenesis of P. japonicus. Since no available germ cell lines of crustacean, we can’t conduct the experiments in vitro and explain the effects more clearly. However, our results are still very important.

KIFC1 functions in the acrosome formation and nuclear morphogenesis during spermiogenesis

The motor protein KIFC1 has been proved to function in the formation of acrosome and morphogenesis of the nucleus in different species from lower animals like invertebrates to mammalians (Table 1). It is related to the nuclear membrane and acrosome in round and elongating spermatids in mammal [56]. In rats, KIFC1 may combine with nuclear factors and contribute to the formation of spermatid acrosome, as well as the nuclear changes, which gives the direct association of a molecular motor with the acrosome biogenesis and nuclear morphogenesis [17]. Wang et al. [32] came up with the initial hypothesis that KIFC1 keep the conserved role during spermiogenesis of cephalopods: participating in acrosome formation and nucleo-morphogenesis. And then, they detected that kifc1 mRNA signal was abundant in spermatid undergoing acrosome formation and dramatic elongation in Octopus tankahkeei. In addition, during nuclear elongation and condensation KIFC1 was completely co-localizated with the cephalopod counterpart of manchette. It suggested that cephalopod KIFC1 took part in nuclear morphogenesis with assisting of the manchette-like perinuclear microtubules to transport specific cargoes [57]. During the swimming crab spermiogenesis such as Portunus trituberculatus and Eriocheir sinensis, several studies also have provided a preliminary evidence for the essential roles of KIFC1 in the acrosome biogenesis and nuclear morphogenesis [33, 37, 58]. However, in the phylum Sipuncula Phascolosoma esculenta, there is something different. Combined the detailed observations of the dynamic distribution of motor protein KIFC1, microtubules and mitochondria during spermiogenesis with the conserved function of KIFC1 in cargo transport and microtubule organization, the researchers proposed that KIFC1 functions in nuclear reshaping and midpiece formation in the spermiogenesis of P. esculenta [35].

The Natantia species of Decapoda is divided into three tribes: Caridea, Penaeid and Stenopodidea. In Caridea, several studies have been reported about KIFC1 participating in the spermiogenesis. Acroframosome (AFS) is a new structure consisted of microtubules, which is identified from M. nipponense. The microtubule bundle of AFS started from middle stage spermatids and localized around the developing acrosome, which assembled in an everted umbrella [38]. KIFC1 was then notarized to transport cargoes such as Golgi vesicles, mitochondria and other cellular components that results in acrosome formation and nuclear shaping with the help of AFS during the spermiogenesis of Exopalaemon modestus [59]. During P. clarkii spermiogenesis, although the AFS was not reported, there is no effect on KIFC1 functions in acrosome formation and nuclear shaping [55]. In this study, the KIFC1 randomly dispersed in the cytoplasm, and co-localized with the microtubules in early stage spermatid of P. japonicus, which is similar with that of P. clarkia, M. nipponense and E. modestus [38, 55, 59]. In the late-stage spermatids, the microtubules exhibited characteristic spindle morphology around the nucleus, while the signal of KIFC1 in this phase presented a relative reduction. In mature sperm from the spermatophore, the nucleus, acrosome and spike were clearly observed by phase contrast microscopy. The signal of microtubules and KIFC1 was completely separated. KIFC1 was localized mainly in the acrosome, but the microtubules were close to the nucleus in the opposite end of acrosome. This is a novel phenomenon, which has not reported in other species. We proposed that since sperm stored in spermatophore were completely mature, and what they need to do is just waiting for fertilization, the motor protein KIFC1 and microtubules have finished their mission. Owing to the limit of the interference duration and survival time after injection which are the two main factors influenced the outcome of this study, it is hardly for us to observe the morphological abnormality and to obtain the direct evidence for more specific function of KIFC1 in the acrosome formation and nuclear morphogenesis.

Animals and sample preparation

The adult male Kuruma shrimp were purchased from the aquatic market of Luojia Village in Hangzhou city (Zhejiang, China). We quickly sacrificed these shrimps and dissected to collect the heart, hepatopancreas, muscle, gill, epididymis, spermatophore and testis on ice. The detached heart, hepatopancreas, muscle, gill were put into liquid nitrogen immediately and then preserved at −80°C for RNA and protein extraction. The detached tissues of epididymis, spermatophore and testis were saved in two parts. One part was stored in –80°C for RNA and protein extraction. The other part was fixed in 4% paraformaldehyde (PFA) in phosphate buffered saline (PBS, 0.1 M, pH 7.4) at 4°C overnight, then dehydrated in 20% sucrose solution in PBS for 8 h and embedded with O.C.T. for immunofluorescence. These samples were stored at −40°C.

RNA extraction and reverse transcription

Total RNA of the heart, hepatopancreas, muscle, gill, epididymis, spermatophore and testis were extracted according to the manufacturer's instructions of Trizol reagent (Takara, Dalian, China). All samples were dissolved and homogenized within the Trizol reagent by a homogenizer. The homogenate was sequentially dealt with chloroform, isopropanol and 75% ethanol to make the RNA precipitated. Finally, we dissolved the precipitated RNA from each tissue in 100 μl DEPC-H₂O and measured its concentration by micro-spectrophotometer (Nano-100, Allsheng). All total RNA samples were stored at −80°C.

The PrimeScript® RT reagent Kit (Takara) was used to reverse transcript the total RNA into cDNA. All samples were stored at −20°C for semi-quantitative PCR analysis of KIFC1 mRNA expression in different tissues. In addition, the testis cDNA was used for kifc1 gene cloning. The Smart RACE cDNA Amplification Kit (CloneTech) and the 3′-Full RACE Core Set with PrimeScript™ RTase (Takara) were used for 5′ and 3′ rapid-amplification of cDNA ends (RACE) reverse transcription, respectively.

Full-length of kifc1 cloning

It was reported that kinesin protein KIFC1 is conserved in its C-terminal. Thus, we try to design two pairs of primers (Table 2) based on the conserved region of kifc1 from Eriocheir sinensis by using the Primer Premier 5 software to get the conserved sequence of the shrimp kifc1 in the testis. All the primers were synthesized by the Shanghai Sangon Biological Engineering Technology Company (China). A cDNA fragment of kifc1 was obtained by Nested Touchdown PCR using a Mygene Series Peltier Thermal Cycler (Hangzhou, China). The program ran as follows: 94°C for 5 min, 10 cycles of the touchdown program (94°C for 30 s, 55°C for 30 s, 72°C for 1.5 min, followed by 0.5°C decrease of the annealing temperature per cycle), followed by 25 cycles (94°C for 30 s, 50°C for 30 s, and 72°C for 1.5 min), and 10 min at 72°C for the final extension. The PCR products were then determined and separated by agarose electrophoresis. SanPrep column DNA gel extraction kit (Sangon Biotech) was used to purify the desired bands. Eventually, the purified fragment was linked into PMD19-T-vector (Takara), propagated in Escherichia coli DH5α (Takara) and then sent to the Sangon Biological Engineering Technology Company for sequencing.

Based on this determined sequence, we designed two forward primers (3'F1 and 3'F2) for 3′ RACE and one reverse primer (5'R1) for 5’ RACE using the Primer Premier 5 software. Then we performed the Nested Touchdown PCR with the reverse primers from the 3′-Full RACE Core Set with PrimeScript™ RTase to abtain the 3′ RACE, and with the forward primer from the Smart RACE cDNA Amplification Kit (CloneTech) to get the 5′ RACE. We ran the 3’ RACE programs as follows: 94°C for 5 min, 10 cycles of the touchdown program (94°C for 30 s, 62°C for 30 s, 72°C for 0.5 min, followed by 0.5°C decrease of the annealing temperature per cycle), followed by 25 cycles (94°C for 30 s, 57°C for 30 s, and 72°C for 0.5 min), and 10 min at 72°C for the final extension. The 5′ RACE programs was as follows: 94°C for 5 min, 10 cycles of the touchdown program (94°C for 30 s, 68°C for 30 s, 72°C for 0.5 min, followed by 0.5°C decrease of the annealing temperature per cycle), followed by 25 cycles (94°C for 30 s, 63°C for 30 s, and 72°C for 0.5 min), and 10 min at 72°C for the final extension. The aimed PCR products were purified, amplified and sequenced by Sangon Biological Engineering Technology Company as described above.

Sequence analysis and phylogenetic analysis

The full-length KIFC1 cDNA sequence was checked and assembled by Seqman (DNASTAR, Inc.). The amino acid sequence was translated by the online ExPASy translate tool (http://web.expasy.org/translate/). Multiple sequence alignments were performed with Vector NTI10 (Invitrogen). The secondary structure of KIFC1 protein was predicted with an online tool PSIPRED Workbench (http://bioinf.cs.ucl.ac.uk/psipred/). The 3-D structure of KIFC1 was predicted with an online server I-TASSER (http://zhanglab.ccmb.med.umich.edu/I-TASSER). The phylogenetic tree, neighbor-joining (NJ) method, was constructed according to the deduced amino acid sequence using MEGA 6. The aa sequence of KIFC1 homologues used in this study was downloaded from NCBI. Their Genbank accession numbers were as follow: Homo sapiens (NP_002254.2), Bos taurus (NP_001095406.1), Mus musculus (NP_001182227.1), Rattus norvegicus (NP_001005878.1), Gallus gallus (NP_001075167.1), Pseudonaja textilis (XP_026580999.1), Python bivittatus (XP_007442391.1), P. chinensis (AFP33411.1), Alligator sinensis (XP_006032359.1), Danio rerio (NP_571281.2), Salmo salar (ABQ59663.1), Larimichthys crocea (ALP83460.1), Micropterus salmoides (AKS36882.1), E. sinensis (ADJ19048.1), Portunus trituberculatus (AKS36885.1), P. clarkii (AJF36162.1), M. nipponense (AFP33456.1), Palaemon modestus (AIN36847.1), and M. rosenbergii (AFO63546.1).

Semi-quantitative PCR analysis of KIFC1 mRNA expression

A pair of primers (KIFC1-RTF1, KIFC1-RTR1) (Table 2) were designed for analyzing the expression of KIFC1 mRNA in different tissues: heart, hepatopancreas, muscle, gill, seminiferous duct, spermatophore and testis. A pair of primers (actin-F1, actin-R1) (Table 2) was used to amplify a β-actin cDNA fragment as the control. The PCR program was run as follows: initial incubation at 94°C for 5 min; 35 cycles in a normal program: 94°C for 5 s, 60°C for 10 s, and 72°C for 15 s; 72°C for 5 min for the final extension. The software Image J was used to analyze the expression of KIFC1 mRNA in different tissues.

Western blot

All tissues were homogenized in RIPA Lysis Buffer (Beyotime) with 1% protease inhibitors. The total protein was separated by 10% SDS-PAGE gels and then transferred to PVDF membranes (Millipore) by 200 mA, 90 min. After 2 h blocked with 5% non-fat milk in PBST buffer (0.1% Tween-20 in PBS, pH 7.2-7.4), the membranes were incubated overnight in primary antibodies at 4 °C (KIFC1 rabbit polyclonal antibody (1:500) designed and prepared by our lab, anti-ACTB rabbit polyclonal antibody (1:2000, BBI)). We washed the membranes in PBST (three times, 15min/time). Then the membranes were incubated with secondary goat-anti-rabbit HRP-conjugated antibody (1:4000, Beyotime) for 1h. After washing three times with PBST, an enhanced chemiluminescent kit (Beyotime) was used to examine the protein blots by chemiluminescence imaging. β-actin was used as control.

Immunofluorescence

The samples were taken out from -40°C and cut into 8 μm sections with Cryostat microtome (Thermo Scientific, HM525 NX). The sections were permeabilized with 0.25% Triton X-100 in PBS for 15 min. Then, 5% BSA in PBST buffer was used to block the sections for 1 h. After that, the sections were incubated with KIFC1 rabbit polyclonal antibody (1:100) overnight at 4 °C. The negative control sections were incubated in 5% BSA without primary antibody. They were washed in PBST for 45 min (3 times, 15 min/time). Subsequently, the sections were incubated with Secondary Alexa Fluor 555-conjugated donkey-anti-rabbit antibody (1:500, Beyotime) and anti-α-Tubulin-FITC mouse monoclonal antibody (1:100, Sigma) at room temperature for 1 h. The negative control sections were incubated with just Secondary Alexa Fluor 555-conjugated donkey-anti-rabbit antibody (1:500, Beyotime). After washed 3 times like before, the sections were incubated with DAPI (Beyotime) for 5 min to stain the nucleus. Finally, the sections were mounted with Antifade Mounting Medium (Beyotime) and observed immediately with a confocal laser scanning microscope (Carl Zeiss, CLSM 710).

RNA interference (RNAi)

Two pairs of specific primers of kifc1 and one pair of EGFP negative control primers (Table 2) with restriction sites SpeI/ XhoI were designed by the Primer Premier 5 software. We cloned the target sequences into L4440 plasmid with two T7 promoters. The recombinant plasmid was used as template to synthesize the dsRNA by T7 RNAi Transcription kit (Vazyme). We examined the quality and concentration of dsRNA by 1% agarose gel electrophoresis and a micro-spectrophotometer (Nano-100, Allsheng), respectively.

The active adult male shrimps were selected for the RNA interference experiment. All selected shrimps were divided into four groups (Normal group, dsEGFP group, dsKIFC1-1 group and dsKIFC1-2 group, 12 individuals in each group) and cultured under a stationary temperature of 18°C in separated boxes. After the shrimps were adapted to the environment of the laboratory for 1 week, in vivo injections of dsRNA were given. Each shrimp was injected with 20μg dsRNA in 20 μl RNase-free water every three days, and lasted for two weeks. There was no treatment in normal group.

TUNEL assay

The One Step TUNEL Apoptosis Assay Kit (Beyotime) was used for TUNEL assays. The prepared sections were fixed in 4% PFA for 30 min, and then permeabilized in 0.5% Triton X-100 in PBS for 5 min. After washed in PBS for two times, they were incubated with TUNEL reaction mixture for 1 h at 37 °C away from light, washed for three times in PBS, then stained with DAPI (Beyotime) for 5 min. Finally, the sections were mounted with Antifade Mounting Medium (Beyotime) and observed immediately with a confocal laser scanning microscope (Carl Zeiss, CLSM 710).

Statistical analysis

Statistical analyses were performed by the GraphPad Prism 7.0 software. The t-test was used to determine the difference between two groups. Statistical significance was presented as *p < 0.05; **p < 0.01; ***p < 0.001, and no significance (ns). We indicated the results as means ± SEM.