Close inbreeding is known for a variety of small mammal species for which a high probability of mortality during dispersal makes helping and delayed maturation a relatively secure fitness option. Prolonged inbreeding, however, is usually associated with lowered fitness, and it has been shown that most highly inbred small mammals and social insects have inbreeding-avoidance mechanisms that promote some degree of outbreeding. However, previous field and laboratory research on the naked mole-rat (Heterocephalus glaber) suggested that this cooperatively breeding rodent is highly inbred, with new colonies forming by fission. Here we report the discovery of a dispersal phenotype that may occasionally promote outbreeding in naked mole-rats. These dispersers are morphologically, physiologically and behaviourally distinct from other colony members. They are laden with fat, exhibit elevated levels of luteinizing hormone, have a strong urge to disperse, and only solicit matings with non-colony members. These findings suggest that, although rare, a dispersive morph exists within naked mole-rat colonies.

The African mole-rats (family Bathyergidae) are subterranean hystricomorph rodents occurring in a variety of habitats and displaying levels of sociality which range from solitary to eusocial, making them a unique mammalian taxonomic group to test ecological in£uences on sociality. Here, we use an extensive DNA-based phylogeny and comparative analysis to investigate the relationship between ecology, sociality and evolution within the family. Mitochondrial cytochrome-b and 12s rRNA trees reveal that the solitary species are monophyletic when compared to the social species. The naked mole-rat (Heterocephalus glaber) is ancestral and divergent from the Damaraland mole-rat (Cryptomys damarensis), supporting previous ¢ndings that have suggested the multiple evolution of eusociality within the family. The Cryptomys genus is speciesrich and contains taxa exhibiting di¡erent levels of sociality, which can be divided into two distinct clades. A total of seven independent comparisons were generated within the phylogeny, and three ecological variables were signi¢cantly correlated with social group size: geophyte density (p50.05), mean months per year of rainfall greater than 25 mm (p50.001), and the coe¤cient of rainfall variation (p 0.001). These results support the food^aridity hypothesis for the evolution of highly social cooperative behaviour in the Bathyergidae, and are consistent with the current theoretical framework for skew theory.

African mole-rats are subterranean Hystricomorph rodents, distributed widely throughout sub-Saharan Africa, and displaying a range of social and reproductive strategies from solitary dwelling to the 'insect-like' sociality of the naked mole-rat, Heterocephalus glaber. Both molecular systematic studies of Rodentia and the fossil record of bathyergids indicate an ancient origin for the family. This study uses an extensive molecular phylogeny and mitochondrial cytochrome b and 12s rRNA molecular clocks to examine in detail the divergence times, and patterns of speciation of the five extant genera in the context of rift valley formation in Africa. Based on a value of 40-48 million years ago (Myr) for the basal divergence of the family (Heterocephalus), we estimate divergence times of 32-40 Myr for Heliophobius, 20-26 Myr for Georychus/Bathyergus and 12-17 Myr for Cryptomys, the most speciose genus. While early divergences may have been independent of rifting, patterns of distribution of later lineages may have been influenced directly by physical barriers imposed by the formation of the Kenya and Western Rift, and indirectly by accompanying climatic and vegetative changes. Rates of chromosomal evolution and speciation appear to vary markedly within the family. In particular, the genus Cryptomys appears to have undergone an extensive radiation and shows the widest geographical distribution. Of the two distinct clades within this genus, one exhibits considerable karyotypic variation while the other does not, despite comparatively high levels of sequence divergence between some taxa. These different patterns of speciation observed both within the family and within the genus Cryptomys may have been a result of environmental changes associated with rifting.

To investigate the endocrine cause of reproductive suppression in nonbreeding female naked mole-rats, animals from 35 colonies were studied in captivity. Urinary and plasma progesterone concentrations were elevated in pregnant females (urine: 10.0-148.4 ng/mg Cr, 27 samples from 8 females; plasma: 3.6-30.0 ng/ml, 5 samples from 5 females; Days 21-40 of pregnancy) and cyclic breeding females (urine: 0.5-97.8 ng/mg Cr, 146 samples from 7 females; plasma: less than 1.0-35.4 ng/ml, 25 samples from 7 females). The latter group showed cyclic patterns of urinary progesterone, indicating a mean ovarian cycle length of 34.4 +/- 1.6 days (mean +/- s.e.m.) with a follicular phase of 6.0 +/- 0.6 days and a luteal phase of 27.5 +/- 1.3 days (19 cycles from 9 breeding females). In non-breeding females urinary and plasma progesterone values were undetectable (urine: less than 0.5 ng/mg Cr, 232 samples from 64 females; plasma: less than 1.0 ng/ml, 7 samples from 6 females). Breeding females had higher (P less than 0.001) plasma LH concentrations (3.0 +/- 0.2 mi.u./ml, 73 samples from 24 females) than did non-breeding females (1.6 +/- 0.1 mi.u./ml, 57 samples from 44 females). Urinary and plasma progesterone concentrations in non-breeding females from wild colonies situated near Mtito Andei, Kenya, were either below the assay sensitivity limit (urine: less than 0.5 ng/mg Cr, 11 females from 2 colonies; plasma: less than 1.0 ng/ml, 25 females from 4 colonies), or very low (plasma: 1.6 +/- 0.6 ng/ml, 15 females from 4 colonies). In captivity, non-breeding females removed from their colonies (i.e. the dominant breeding female) and either paired directly with a non-breeding male (N = 2), or removed and housed singly for 6 weeks before pairing with a non-breeding male (N = 5) may develop a perforate vagina for the first time in as little as 7 days. Urinary progesterone concentrations rose above 2.0 ng/mg Cr (indicative of a luteal phase) for the first time 8.0 +/- 1.9 days after being separated. These results suggest that ovulation is suppressed in subordinate non-breeding female naked mole-rats in captive and wild colonies, and show that plasma LH concentrations are significantly lower in these non-breeding females. This reproductive block in non-breeding females is readily reversible if the social factors suppressing reproduction are removed.

Aging is a major risk factor for progression of liver diseases to hepatocellular carcinoma (HCC). Cellular senescence contributes to age-related tissue dysfunction, but the epigenetic basis underlying drug-induced senescence remains unclear.macroH2A1, a variant of histone H2A, is a marker of senescence-associated heterochromatic foci that synergizes with DNA methylation to silence tumor-suppressor genes in human fibroblasts. In this study, we investigated the relationship between macroH2A1 splice variants, macroH2A1.1 and macroH2A1.2, and liver carcinogenesis. We found that protein levels of both macroH2A1 isoforms were increased in the livers of very elderly rodents and humans, and were robust immunohistochemical markers of human cirrhosis and HCC. In response to the chemotherapeutic and DNA-demethylating agent 5-aza-deoxycytidine (5-aza-dC), transgenic expression of macroH2A1 isoforms in HCC cell lines prevented the emergence of a senescent-like phenotype and induced synergistic global DNA hypomethylation. Conversely, macroH2A1 depletion amplified the antiproliferative effects of 5-aza-dC in HCC cells, but failed to enhance senescence. Senescence-associated secretory phenotype and whole-transcriptome analyses implicated the p38 MAPK/IL8 pathway in mediating macroH2A1-dependent escape of HCC cells from chemotherapy-induced senescence. Furthermore, chromatin immunoprecipitation sequencing revealed that this hepatic antisenescence state also required active transcription that could not be attributed to genomic occupancy of these histones. Collectively, our findings reveal a new mechanism by which drug-induced senescence is epigenetically regulated by macroH2A1 and DNA methylation and suggest macroH2A1 as a novel biomarker of hepatic senescence that could potentially predict prognosis and disease progression.