Diet-Induced Modification of the Sperm Epigenome Programs

Trends in Endocrinology & Metabolism

Review Diet-Induced Modiﬁcation of the Sperm Epigenome Programs Metabolism and Behavior

Carina Bodden,1,* Anthony J. Hannan,1,2,* and Amy C. Reichelt1,3,4

Globally, obesity has reached epidemic proportions. The rapidly increasing numbers of over- Highlights weight people can be traced back to overconsumption of energy-dense, poor-quality foods as Epigenetic programming is a well as physical inactivity. This development has far-reaching and costly implications. Not only meaningful nongenetic process for is obesity associated with serious physiological and psychological complications, but mounting conveying information from one evidence also indicates a ripple effect through generations via epigenetic changes. Parental generation to the next. However, obesity could induce intergenerational and transgenerational changes in metabolic and brain paternal epigenetic inheritance has function of the offspring. Most research has focused on maternal epigenetic and gestational ef- received scant attention. fects; however, paternal contributions are likely to be substantial. We focus on the latest ad- Paternal obesity, as a result of an vances in understanding the mechanisms of epigenetic inheritance of obesity-evoked metabolic excessive consumption of calorie and neurobiological changes through the paternal germline that predict wide-ranging dense, poor-quality foods, can consequences for the following generation(s). predict adverse health consequences for the offspring.

Junk Food, the Obesity Epidemic, and Transgenerational Effects of Diet Rodent models of obesity have Unhealthy diets can have negative long-term consequences on subsequent generations. This alarm- revealed diet-induced alterations ing conclusion is largely the result of recent transgenerational studies in rodents [1,2]. This finding has in sperm and seminal plasma that represent paternal transmission particular translational relevance to humans because obesity (see Glossary)inyouthandadultsof of epigenetic markers. These child-bearing age is dramatically increasing [3]. Globally, the number of obese children has risen may influence the development, tenfold in the past four decades [4]. Childhood obesity, especially during puberty, is associated behavior, and brain function of the with obesity in adulthood [5] as well as with obesity in either parent [6], forming a cycle of intergen- offspring. In particular, sperm small erational transmission (and potentially transgenerational transmission), escalating the prevalence of and long noncoding RNAs appear obesity. The transmission of a predisposition to obesity across generations may reflect classic genetic to be prime candidates for medi- inheritance of risk alleles (e.g., [7]) from parents to offspring, such as polymorphisms in the fat mass ating dietary effects through the and obesity-associated (FTO) gene or rare congenital variants (i.e., leptin receptor deficiency). paternal line. However, the absence of evidence for substantial changes to the human genome over the past Further research will be necessary four decades means that genetics alone cannot explain the rapidly increasing rate of obesity. Instead, to define the mechanisms underly- recent correlative evidence suggests that epigenetic information plays an important role in the trans- ing diet-induced alterations in mission of obesity through generations [8]. paternal sperm and offspring phenotype. Epigenetics can be viewed as a point of crosstalk between the genome and the environment, enabling environmental stimuli, including diet, to bring about stable molecular and behavioral modifications that can be transmitted to offspring [9]. Epigenetic modifications can mediate long-lasting physiological, neurobiological, and behavioral changes in offspring [10]. However, the molecular mechanisms underlying such transgenerational transmission from parents to offspring – ranging from embryonic development to subsequent offspring behavior, health, and disease risk – are far from understood. 1The Florey Institute of Neuroscience and Mental Health, University of Melbourne, Despitetheequalprevalenceofobesityinbothsexes[11], the vast majority of research has focused Melbourne, 3010 VIC, Australia on maternal nutritional influences on offspring during gestation and lactation. However, the impact of 2Department of Anatomy and Neuroscience, University of Melbourne, paternal diet on offspring has received limited attention [12]. Notably, recent evidence indicates that, Melbourne, 3010 VIC, Australia in addition to stress exposure and drug use, diet and obesity status are crucial paternal preconcep- 3BrainsCAN, Western Interdisciplinary tion environmental factors associated with altered phenotypic outcomes – including disease suscep- Research Building, Western University, tibility – in offspring [13](Figure 1, Key Figure). Epidemiologic studies have demonstrated a correla- London, ON, Canada tion between paternal nutritional status and child health outcomes (reviewed in [14]). Importantly, 4Robarts Research Institute, Western University, London, N6A 3K7 ON, Canada analogous findings have been made using animal models of obesity in the laboratory, which provide *Correspondence: useful tools to identify phenotypic outcomes, the underlying mechanisms, and potential interven- carina.bodden@florey.edu.au, tions [14]. anthony.hannan@florey.edu.au

Key Figure Direct and Intergenerational Epigenetic Effects of an Obesogenic Diet on Male Individuals and Their Progeny

Figure 1. This schematic figure is a simplified illustration of the wide range of effects of an energy-rich, poor-quality diet as well as of the proposed mechanisms by which it can have intergenerational long-term consequences for the offspring. Energy-dense foods replete with saturated fats and refined sugars cause a variety of symptoms in males that affect their gut microbiome, hormone levels, and sperm health. RNA species contained in sperm and seminal fluid are influenced by an unhealthy diet, leading to abnormalities in fertilization and embryonic development, presumably via different epigenetic processes, that predict several phenotypic changes in the offspring, including metabolic, reproductive, and cognitive dysfunctions. Abbreviations: lncRNA, long noncoding RNA; M, DNA methylation

This review will synthesize recent research ﬁndings regarding the biological mechanisms that transmit effects of paternal diet and obesity to the next generation, and the consequences for subsequent generation(s).

How Paternal Diet and Obesity Physiologically Impact on the Metabolic Profiles of Offspring Recent findings from human and rodent studies indicate that the periconceptional phase, encom- passing the stages of gamete maturation, fertilization, and early embryogenesis, is particularly sensitive to environmental influences, including unhealthy high-fat diets (HFDs) [15]. Hence, external stimuli may be capable of causing long-term phenotypic changes through epigenetic, cellular, and physiological mechanisms following either perturbation or adaptive responses [16].

HFD feeding of male mice has recently been demonstrated to modify the sperm epigenome, with consequences for the offspring [17]. DNA methylation and small noncoding RNA (sncRNA) signatures showed

132 Trends in Endocrinology & Metabolism, February 2020, Vol. 31, No. 2 Trends in Endocrinology & Metabolism

great similarities in the epigenome of spermatozoa from HFD-fed fathers and their male offspring. Strik- ingly, transfer of sperm RNA from HFD-fed males into naı¨ve one-cell embryos alone was sufficient to induce Glossary a metabolic phenotype conducive to the development of obesity in the resulting F1 generation [18]. Both Blood–brain barrier:aselective male and female offspring displayed higher body weights than controls, which were either similar to or barrier formed from endothelial cells adjoined with tight junctions higher than those of animals raised on a HFD [18]. Consequently, transfer of nongenetic (i.e., epigenetic) that allows specific molecules information must be involved, with RNA being one of the confirmed vectors. from the blood to permeate by either diffusion or active transport In the following sections we review potential paternal mechanisms of nongenetic information transfer through into the cerebral vascu- that can alter offspring development, including sperm quality and composition, seminal plasma lature or cerebrospinal fluid. Blood–testis barrier:aprotective composition, DNA integrity, and epigenetic alterations to DNA, histones, protamines,andRNAs, semipermeable barrier, similar to as well as the multifaceted consequences for the metabolism and behavior of the offspring [2,19–21]. that of the blood–brain barrier, between blood vessels and the From Mice to Men – What We Can Learn from Rodent Models of Obesity seminiferous tubules where spermatozoa are generated. Because human studies can be highly confounded by lifestyle factors such as smoking and activity Epigenetics: structural adaptation levels [22], animal models have been instrumental in circumventing these limitations and to examine of chromosomal regions that reg- the specific effects of paternal dietary patterns. In rodent models, obesity can be generated by ister, signal, or perpetuate altered providing diets with a high proportion of saturated fats, sugars, or both. Feeding male mice a HFD activity states [9], including DNA methylation, histone post-trans- has been associated with obesity-related metabolic phenotypes such as glucose intolerance and in- lational modifications, mRNA sulinresistance[23]. Hormonal changes extend to decreased testosterone levels and reduced post-transcriptional changes, and androgen receptor gene expression [24]. Dysregulated levels of testosterone can disrupt the regulation by noncoding RNAs. blood–testis barrier during spermatogenesis [24]. This condition likely contributes to the decreased Exosomes: small membrane vesicles of endocytic origin that fertility that has been observed in male mice on an obesogenic diet [24]aswellasinobesemen[25]. transport RNAs and proteins both within and between cells. Fast Food Predicts Slow and Stressed Sperm F0 generation: the initial (i.e., parental) generation, sometimes Obesity-evoked disruptions to the blood–testis barrier may contribute to the dramatically increased also referred to as the P rates of abnormality in basic parameters of sperm health, such as critically reduced motility, concen- generation. tration, ejaculate volume, and altered morphology, that have been observed in both overweight F1 generation: the first filial gen- humans [25,26] and in rodent models of obesity [27]. In addition, similar to findings in obese men eration, namely the first set of offspring from controlled or [26], a HFD in male mice has been linked to elevated production of intracellular reactive oxygen spe- observed reproduction of F0 par- cies (ROS), oxidative stress-induced sperm DNA damage, and reduced sperm mitochondrial function ents. The F1 generation can [21,28], These abnormalities indicate the deleterious effects of HFD on reproductive parameters. reproduce to create the F2 gen- Accordingly, it has been confirmed that a lower number of sperm from HFD-fed mice bind to each eration, and so forth. Gut–brain axis: the bidirectional oocyte, resulting in reduced fertilization rates [21]. axis of communication between the central and the enteric ner- Notably, the type of fatty acids in the diet is decisive for both sperm quality and adipose accumulation vous systems that is proposed to [29,30]. Feeding rats a lard-based diet rich in saturated fatty acids reduced the number of normal be a crucial link between the sperm cells and daily sperm production in comparison with a plant-based HFD, although overall emotional and cognitive centers of the brain and peripheral intes- caloric intake did not differ [29]. Interestingly, supplementation of a high-fat and high-sugar diet tinal functions. (HFHS) with olive oil or krill oil [rich sources of omega-3 (u-3) fatty acids] counteracted the negative Gut microbiome: the ensemble of effects on sperm quality and function. Adding krill oil also reversed the increased adiposity caused by all the microbial entities, genes, HFHS consumption [30]. proteins, and metabolic products within the gastrointestinal system at a given time. In summary, excessive consumption of poor-quality food replete with saturated fats induces oxidative Higher-order cognition:special- stress and sperm DNA damage, leading to decreased fertility. Crucially, evidence indicates that ized cognitive functions that are feeding male mice an unhealthy HFD provokes subfertility in the male and female offspring for at least necessary to interact dynamically two generations [31], an alarming finding that warrants further investigation. Whether paternal within the ever-changing environment. Incorporates cognitive obesity in humans also affects the reproductive health of children and grandchildren in humans is domains that include decision currently not known. making and inhibitory control. Imprinting: epigenetic marks driven by DNA methylation that Diet-Induced Alterations in Seminal Plasma lead to parent-of-origin-specific Recent studies have indicated that the proteomic profile of seminal plasma (Box 1) derived from gene expression; in other words, obese men is significantly altered compared with lean men, and shows markers of increased inflam- some genes are expressed only mation, oxidative stress, and apoptotic mechanisms [32]. Not only do these modifications resemble when they are inherited from the father, whereas others are only obesity-induced alterations in sperm health parameters, but they are also likely to contribute to

Trends in Endocrinology & Metabolism, February 2020, Vol. 31, No. 2 133 Trends in Endocrinology & Metabolism

sperm function via exosomes. Because seminal plasma also possesses its own microbiome [33], of active when inherited from the which the microbial composition is sensitive to even short-term HFD exposure [34], transfer of altered mother. Imprinted genes show RNA cargoes and proteins to sperm could convey information about paternal dietary patterns and resistance to epigenetic re- metabolic status to the oocyte [34]. programming during early embryogenesis. Intergenerational transmission: in vitro Given that a seminal plasma-free environment is the standard for fertilization (IVF) as well as transmission of environment- for intracytoplasmic sperm injection (ICSI), these findings are of great importance and challenge induced changes to subsequent prevailing fertilization techniques, which could be impacting on the health status and disease suscep- generations. For example, when a tibility of offspring (apart from the consequences of epigenetic processes related to the IVF/ICSI tech- pregnant female (F0)isexposedto an environmental stimulus (e.g., nique itself). Supporting evidence comes from a series of studies by Belva and colleagues, who nutrition) and effects are observed compared body composition and sperm of adolescents or young adults, respectively, born after in the F1 offspring and F2 grand- either IVF or spontaneous conception [35,36]. Indeed, adolescents born after ICSI showed increased offspring. Implies that the effects adiposity [35], and ICSI was associated with reduced sperm health parameters in young adult men, of exposure to environmental stimuli are transmitted via devel- independently of paternal semen parameters [36]. Although the authors [35] could exclude an effect oping germ cells. More precisely, of maternal body mass index (BMI), no data are available on the BMIs of the fathers, therefore a the altered germline (sperm or parental effect cannot be completely ruled out. egg) epigenome can impact on the epigenome and transcriptome of the stem cells of the embryo Influence of Diet and Obesity on the Sperm Epigenome and Epigenetic that can in turn influence the epi- Inheritance genome and transcriptomes of all derived somatic cells. The sperm epigenome has recently started to draw increasing attention in terms of its contribution Intracytoplasmic sperm injection to embryo development [46]. The previous underappreciation was based on the fact that during (ICSI): a more invasive option than spermiogenesis the majority of nucleosomes – as well as most epigenetic information – is removed conventional in vitro fertilization (IVF) that involves direct injection of a single sperm cell into the oocyte cytoplasm, which can be Box1.SeminalPlasma–MoreThanaFluid successful even with poor semen Seminal plasma is a highly complex composite secretion arising from different sources, including the testes, characteristics or testicular (i.e., immature) sperm. epididymes, seminal vesicles, and accessory glands, that contains proteins, amino acids, enzymes, carbohy- In vitro fertilization (IVF):apro- drates, lipids, mRNAs, and miRNAs (reviewed in [37]). Although seminal plasma was long thought to serve cedure in which an oocyte is the sole purpose of delivering spermatozoa to the oocyte, seminal fluid also contains signaling molecules, fertilized outside the body. The such as several cytokines and growth factors, that can prompt gene expression and initiate essential fertilized egg is allowed to grow in immunological responses in the maternal reproductive tract [38,39]. Furthermore, seminal plasma comprises a protected laboratory environ- a variety of proteins that play an important role in sperm maturation, motility, protection, and oocyte ment before being transferred binding [40]. These sperm functions are modulated via extracellular vesicles such as exosomes, that contain into the uterus. proteins as well as RNAs, that bind to the spermatozoa surface [40]. Epididymosomes represent a Mesocorticolimbic dopamine subgroup of exosomes whose protein composition undergoes changes during the transit through the system: the reward and rein- forcement pathway of the brain epididymis. Because there is evidence indicating a role in gametogenesis, post-testicular sperm maturation, that includes several inter- fertilization, and embryo development, epididymosomes may be involved in programming offspring long- connected brain areas. It involves term health [41]. Importantly, epididymosomes have been found to fuse with and transfer RNA cargoes the signaling molecule dopamine from somatic cells to germ cells [42], likely an important mechanism for germline-mediated epigenetic and provides salience to internal transmission. and external stimuli. It is a central factor in the development of Strikingly, experimental studies have shown that lack of seminal plasma during conception in mice not only addiction. causes a delay in preimplantation embryo development but also promotes the development of obesity and Nucleosome:asegmentofDNA glucose intolerance in the adult offspring, highlighting the importance of seminal plasma in metabolic that is wrapped around a core of programming [43]. Further evidence from human IVF studies suggests that exposure to seminal plasma has eight histone proteins. positive effects on endometrial function and the maternal immune system, thereby supporting implantation Obesity: a term used to [44]. denote abnormal or excessive fat accumulation that presents a risk The majority of studies to date have focused on either sperm-specific epigenetic changes or seminal plasma- to health. The body mass index mediated programming, but few in vivo experiments have investigated both variables. Because animal (BMI) – a person’s weight (in kilo- models provide the option to disentangle individual variables involved in a specific process, Watkins and grams) divided by the square of colleagues [45] characterized both sperm- and seminal plasma-mediated effects of paternal diet on offspring his or her height (in meters) – represents an approximate pop- in mice by combining artificial insemination with mating involving vasectomized males, and demonstrated ulation measure of obesity. A diet-dependent paternal programming effects through sperm (epigenetic status) as well as seminal plasma person with a BMI of 30 or more is (preimplantation uterine environment). They identified sperm- and seminal plasma-specific programming ef- considered obese, whereas a fects of a paternal low-protein diet [45], clearly indicating the relevance of seminal plasma for embryonic person with a BMI of R25 is programming. considered overweight [4].

134 Trends in Endocrinology & Metabolism, February 2020, Vol. 31, No. 2 Trends in Endocrinology & Metabolism

and replaced by small basic proteins, called protamines, resulting in a highly compact genome that Protamines: in spermatozoa, the is essential for fertility and genome integrity [46]. An obesogenic diet has been associated with al- majority of genomic DNA is terations in the acetylation of spermatids, a process that is important for appropriate histone wrapped around protamines, al- replacement by protamines. The result is incorrect histone-to-protamine exchange leading lowing denser packaging of DNA than with histones. After fertiliza- to less-condensed chromatin which is vulnerable to oxidative stress and DNA damage (reviewed tion, paternal protamines are rein [47]). placed by maternal histones. Spermatogenesis: the process of Recent findings also indicate that the sites of nucleosome retention are not randomly dispersed along gradual transformation of male the chromosomes, but are instead enriched at particular genomic regions that regulate, for example, primordial germ cells into sperm cells, which takes place mainly embryo development, providing the possibility for epigenetic programming in the embryo ([46,48– within the seminiferous tubules. 50], reviewed in [51]). The observed DNA methylation (Box 2), histone, and chromatin features (re- Spermiogenesis: the final differ- viewed in [10]) may also be involved in the intergenerational and transgenerational transmission of entiation and maturation process paternal diet that has been linked to metabolic disorders and subfertility in up to two generations of spermatids into sperm cells. Transgenerational transmission: in several rodent studies [2,17,31,52](Table 1). inheritance of an altered phenotype by generations that were not Accumulating evidence demonstrates that micronutrients such as phytochemicals, minerals, and vi- exposed to the initial signal or tamins can evoke epigenetic modifications of germ cells [47]. Noncoding RNA transcripts are also environment that triggered the increasingly recognized as a source of paternal programming; however, the coding mechanisms change. For example, when the signal is experienced by the male behind RNA structures, RNA–DNA interactions, and other factors remain elusive [53]. Nevertheless, (F0)ornon-pregnantfemale(F0) the phase of sperm maturation during the transit through the epididymis – a tubular structure in which but the effect is observed in the F2 sperm undergo post-testicular maturation – seems to play a crucial role in the modulation of epige- grand-offspring and beyond. netic information. According to a recent study [54], mature sperm cells contain RNA molecules that Thus, transgenerational transmission involves alteration of the were first synthesized in the epididymal epithelium and were then transferred to the maturing sperm germline epigenome. during their post-testicular maturation.

Box2.DNAMethylationinSperm One of the major epigenetic mechanisms controlling gene expression and imprinting is the methylation of cytosine residues in DNA [55]. Recently, de novo DNA methyltransferase DNMT3A was found to play a crucial role in paternal imprinting as well as in spermatogenesis and methylation at several paternally imprinted loci [56]. HFD feeding has been associated with global hypomethylation of germ cell DNA in male mice [2,45]. Notably, global sperm hypomethylation has previously been linked to subfertility in humans [57]. Because a negative correlation between DNA methylation and DNA fragmentation was found [57], it may be speculated that the methylation process is impaired by oxidative stress-related damage of sperm DNA in obese men.

Methylation of genes in the male germline is commonly believed to be largely erased and reprogrammed upon fertilization and during the formation of primordial germ cells; stem cells in the early embryo then undergo re- methylation in a cell type-specific way [58]. However, recent findings from genome-wide studies indicate that sex-specific methylation imprinting is largely resistant to genome-wide demethylation and de novo remethy- lation during embryogenesis [59]. These differentially methylated regions (DMRs), referred to as epimutations, have the potential to affect the development of the next generation [60]. Indeed, several rodent studies have demonstrated an association between the sperm methylation patterns of HFD-fed fathers and sons

[17,48,60,61]. However, although DMRs have been found in the sperm of both F0 fathers and F1 sons, no alter-

ations in these regions were present in F2 grand-offspring adipose tissue, despite changes in gene expression [17]. As shown by de Waal and colleagues [62] in ICSI-derived mice, epimutations at imprinted genes can be

corrected in F1 sperm and are consequently not transmitted to the F2 generation. It should be noted, however,

that DMR changes in F3 offspring have been observed following exposure to toxic substances such as the pesti- cide dichlorodiphenyltrichloroethane (DDT) [63,64]. Whether the intergenerational and transgenerational effects of obesity and hypercaloric diets are similar to those induced by toxins warrants further investigation because other epigenetic and genetic consequences may be evoked by diet.

Hence, although methylation has been studied extensively, the variable nature of epigenetic marks has obscured the exact mechanism of how sperm DNA methylation impacts on the following generation(s). Further research will be necessary to clarify potential additional carriers of molecular information in sperm, such as histone modiﬁcations and RNAs, which can act in trans (i.e., show aspects of resistance to global DNA-methylation erasure), thereby escaping the reprogramming process [10,65].

Trends in Endocrinology & Metabolism, February 2020, Vol. 31, No. 2 135 Trends in Endocrinology & Metabolism

Table 1. Selection of Studies Providing Evidence for Diet-Induced Effects on Male Reproductive Healtha,b

c Species Treatment Effects on F0 Refs

Humans Men BMI >30 Altered expression level of speciﬁc miRNAs, [68] piRNAs, tRFs, and snRNA fragments in spermatozoa Differentially expressed cocaine and amphetamine regulated transcript (regulator of food intake involved in obesity) Numerous differentially methylated genes in motile spermatozoa Weight loss-induced alteration of sperm epigenome

Mice: C57BL/6 HFD: 45% fat Body weight [ [24] Levels of total cholesterol [ Levels of low-density lipoprotein [ Levels of high-density lipoprotein [ Hepatic steatosis [ Fat vacuoles in hepatic cells[ Serum estradiol levels [ Serum testosterone levels Y Serum luteinizing hormone levels Y Percentage of sperm motility and progressive motility Y Fertility (induced pregnancies in females) Y Integrity of the blood–brain barrier Y Androgen receptor expression Y

Humans Men BMI >30 Age, ejaculate volume, sperm concentration and [32] progressive motility: Sperm morphology Y Nonprogressive motility [ Acrosome integrity and mitochondrial activity Y Sperm DNA fragmentation rates [ Inﬂammatory, immune, and complement response [ Over-represented phagocytic protein activity and antioxidant activity in seminal plasma Chaperone concentration, such as HSP90 Y

Rats: HFD: 35% fat Body weight [ [28] Sprague–Dawley Epididymal, inguinal, mesenteric, and retroperitoneal white adipose tissue [ Levels of glucose and insulin in plasma [ Plasma triglyceride levels [ Sperm concentration and motility Y LDH-C4 (lactate dehydrogenase C4) and PDH (pyruvate dehydrogenase) complex activity Y ATP content Y LPO (lipid peroxidation) levels [ Enzymatic activity of respiratory complexes II, III, and IV Y

(Continued on next page)

136 Trends in Endocrinology & Metabolism, February 2020, Vol. 31, No. 2 Trends in Endocrinology & Metabolism

Table 1. Continued

c Species Treatment Effects on F0 Refs

Mice: C57BL/6 HFD: 22% fat Seminal ﬂuid: [34] Corynebacterium spp. [ Acinetobacter johnsonii, Streptophyta, Ammoniphilus spp., Bacillus spp., and Propionibacterium acnes Y Fecal samples: Rikenellaceae [ Bacteroides ovatus, another Bacteroides species, Bilophila, Sutterella spp., Parabacteroides, Biﬁdobacterium longum, Akkermansia muciniphila, an uncharacterized bacterium, and Desulfovibrio spp. Y aTreatments include a HFD in rodents or a BMI of >30 in humans. bKey, [, increase; Y, decrease. cCompared with the control group.

Diet-Evoked RNA Modiﬁcations to Sperm Although mature sperm are transcriptionally and translationally inactive, they contain a range of different RNAs [65]. The presence of RNAs provides the possibility that sperm are capable of delivering paternal information inadditiontothecontainedDNA[66]. Small noncoding RNAs (sncRNAs) are 18–24 bases in length and include repetitive elements, transcription start sites, piwi-interacting RNAs (piRNAs), miRNAs, transfer RNA-derived small RNAs (tsRNAs), and small nuclear RNAs (snRNAs). These sperm-derived RNAs appear to be sensitive to a variety of psychological and physiological factors, including stress exposure and nutrition [67]. Accordingly, obesity in men [68] and mice [23] has been linked to alterations in several subtypes of sperm sncRNAs. Microinjection studies have indicated that sperm sncRNAs are capable of transferring aspects of an acquired phenotype to the next generation, including metabolic and behavioral changes [18,23,69,70]. Mechanistically, this presents opportunities for paternal programming in response to environmental stimuli such as stressors, environmental toxins, and unhealthy diets [69]. The ﬁnding that mature sperm sncRNAs (from post-testicular maturation sperm) appeared to be essential for embryonic development in mice further highlights the crucial importance of sperm RNAs for the offspring [71].

Changes to miRNA Sperm RNA types and quantities fluctuate during sperm maturation. After exiting the testis, sperm maturation continues for several days in the epididymis [42]. During this epididymal transit, substantial modifications to miRNA signatures take place [72]. Structural alterations to epididymal tissue, such as increased fat pad size, could therefore potentially change miRNA signatures [29]. Because miRNAs influence gene expression via mRNAs, their transfer into the oocyte cytoplasm during fertilization [73] could lead to deviations in the epigenetic state of the developing embryo because of reduced target mRNA levels in the zygote [66,74,75]. Therefore, sperm miRNAs have been proposed to act as mediators of paternal programming effects [76]. Indeed, Fullston and colleagues [2]were able to demonstrate diet-induced alterations in both the testicular and sperm transcriptional profiles that were associated with changes in miRNA content [2].

Changes to tsRNA During epididymal transition, sperm gain a high abundance of tsRNAs [42], further emphasizing the potentialimportanceofepididymaltissuemodiﬁcations. Interestingly, substantial changes to murine sperm tsRNAs were observed in response to HFD consumption, indicating increased sensitivity to dietary effects [23]. Moreover, HFD- as well as low-protein diet-induced effects on tsRNAs have been associated with transgenerational transmission of metabolic disorders [23,42,77]. The sequences

Trends in Endocrinology & Metabolism, February 2020, Vol. 31, No. 2 137 Trends in Endocrinology & Metabolism

contained in tsRNAs matched to gene promoter regions associated with transposable elements [53,78]. In addition to eliciting their effects by interactions with DNA, tsRNAs have been shown to act on the translational machinery and on ribosome biogenesis [53], and might therefore inﬂuence embryonic gene expression in various ways.

Changes to lncRNAs Furthermore, long noncoding RNAs (lncRNAs), with a length of >200 nt, have been shown to play a crucial role in spermatogenesis and in paternal epigenetic inheritance of obesity in mice [69,79,80]. These noncoding transcripts can inﬂuence gene expression by modulating the dynamic spatial structure of chromatin during differentiation and development (reviewed in [81]).

Changes to m6ARNA Only recently, HFD feeding of mice has been associated with differential N6-methyladenosine (m6A) RNA modiﬁcation, which is an important post-transcriptional mechanism of gene regulation that is involved in RNA splicing, stabilization, and translation [82]. However, the exact mechanism and consequences for the offspring remain to be deﬁned.

Whether these changes in noncoding RNA proﬁles are adaptive or not, and whether the different epigenetic mechanisms are linked or act independently, still needs to be examined. However, the ability to inject noncoding RNAs into fertilized eggs and investigate offspring phenotypes has pro- vided substantial evidence for their crucial roles in intergenerational and transgenerational epigenetic inheritance [18,23].

Nongenetic Effects of Paternal Diet and Obesity on Progeny Because alterations to sperm and seminal plasma are indeed capable of inﬂuencing the developing organism from as early as the stage of fertilization, the question arises of how these affect the progeny in the short- and long-term. An increasing number of studies indicate paternally transmitted effects of a poor-quality diet on early developmental, metabolic, and behavioral parameters (Figure 2 and Table 2).

Prenatal Development A growing body of evidence from preclinical studies in rodents reveals that the progeny of HFD males undergo abnormal embryonic development, with a reduction in the cleavage rate of zygotes, a signiﬁcant delay in preimplantation development, and reduced implantation, leading to markedly fewer embryos descending from HFD-fed males reaching the blastocyst stage compared with controls [83]. Of those embryos that reached the blastocyst stage, blastocyst cell numbers were decreased in embryos sired by HFD-fed males, which is indicative of reduced viability. Moreover, blastocysts from obese males showed a greater proportion of cells that were apoptotic [83]. Strikingly, recent data from a human fertility center [84] revealed prolonged durations of embryonic cell cycles, with delayed cleavage intervals in the embryos fertilized by obese men compared with those of lean men. Such delays have been associated with a reduced rate of blastocyst expansion and implantation. A similar phenomenon in humans has been observed with regard to advanced paternal age [85]. Mechanisms that could be responsible for this delay include DNA damage in spermatozoa, as noted previously [21,26,86].

Furthermore, small changes in sperm RNAs have the potential to produce long-term metabolic consequences in offspring by regulating embryo mRNA at an early stage through altered gene expression within metabolic pathways [23]. Moreover, the cord blood from newborn offspring of obese fathers showed distinct hypomethylation at the insulin-like growth factor 2 (IGF2) differentially methylated regions (DMRs) [87]. Because changes in the methylation of the IGF2 gene have been linked to chronic adverse health outcomes and an increased risk of developing cancer, the authors suggested that obesity-associated hormone imbalances might mediate incomplete or disorganized methylation of the IGF2 gene during spermatogenesis [87].

138 Trends in Endocrinology & Metabolism, February 2020, Vol. 31, No. 2 Trends in Endocrinology & Metabolism

Figure 2. Range of Intergenerational Effects of a Paternal High-Fat/High-Sugar (HFHS) Diet. Illustrated is the wide-ranging impact that an unhealthy paternal diet may have on offspring physiological, cognitive, and behavioral parameters. Information about the paternal diet is carried to the female oocyte by diverse epigenetic and cellular mechanisms in sperm and seminal plasma, leading to modiﬁcation of early embryogenesis (image created with BioRender, http://biorender.io). For mesocorticolimbic dopamine system see Glossary.

Metabolism Although reduced birth weights in the offspring of obese fathers have been reported in both humans and mice [61,88], a paternal obesogenic diet predicted increased postnatal weight gain in mice [2]. The overall metabolic health of offspring was demonstrated to be distinctly affected by paternal HFD, irrespective of whether an altered glucose homeostasis was present [61]orabsent[2] in the founder males. Female offspring of HFD males displayed metabolic disruptions including reduced glucose tolerance, increased insulin resistance, and adiposity. Furthermore, altered pancreatic function was observed in the form of elevated insulin secretion and reduced pancreatic islet cell size [2,61,89]. Male offspring, however, exhibited altered glucose and insulin sensitivity only later in life. Crucially, most metabolic effects were still present in the subsequent generation, leading to adiposity in grand-offspring (F2 generation) as a transgenerational effect of the diet of the founder males [2].

Trends in Endocrinology & Metabolism, February 2020, Vol. 31, No. 2 139 Trends in Endocrinology & Metabolism

Table 2. Studies Providing Evidence for Diet-Induced Effects on Male Reproductive Health as well as Paternal Intergenerational and/or a,b Transgenerational Inheritance and Effects on the F1 and F2 Generations

c c Species Treatment Effects on F0 Effects on F1 Refs

Mice: C57BL/6J HFD: 60% fat Altered lncRNA and mRNA expression Altered lncRNA and mRNA expression [79] proﬁles in spermatozoa proﬁles in spermatozoa

Similarly expressed transcripts in the F1-HFD

and F0-HFD groups were highly enriched in pathways linked to obesity

Mice: C57BL/6J HFD: 60% fat Body weight [ F1 produced by F0-HFD sperm heads: [23] Glucose intolerance and insulin resistance Impaired glucose tolerance and insulin Larger proportion of tsRNAs showed resistance

differences compared with miRNAs F1 produced by injection of RNAs from F0- HFD sperm into normal zygotes: Impaired GTT: blood glucose and serum insulin levels [ No altered insulin sensitivity

Mice C57BL6/N HFD: 45% fat Glucose AUC in response to insulin Y [89] Insulin-stimulated Akt threonine 308 phosphorylation [ Percent body fat [ Percent lean mass Y

Rats: Sprague– High SFA diet: Body weight [ F1 females: [29] Dawley 60% lard Epididymal fat pad weights [ Birth weight and weight gain [ Normal sperm cells Y Retroperitoneal fat weights [ Daily sperm production Y Fasting glucose levels [ Fasting glucose levels [ Glucose AUCs [ Glucose AUCs in the insulin tolerance test [

Rats: Long–Evans HFD: 35% fat Body weight [ Litter size, litter sex ratio, birth weight = [13] hooded Abdominal and perigonadal fat pads Maternal LG-ABN Y 6 weeks after termination of HFD feeding [ Line crossings in OF [ Brain weight Y (absolute mass comparison [) Entries into open arms of EPM Y Testes weight Y (absolute mass comparison Line crossings in EPM Y Y) Time and percent time in open arms of EPM Time in central area and rearing in OF Y Y Overall line-crosses in OF = Time rearing in EPM [ Time and percent of total time in open arms, Time spent in vicinity of predator odor [ arm entries, time spent rearing in EPM Y Contact with predator odor [ Estrous females preferred control versus Grooming in the presence of predator odor HFD males immediately following diet Y exposure as well as following 20 days of Body weight [ standard chow Brain weight Y Relative weight of abdominal fat pads [ Gonadal fat pads [

Mice: C57BL/6 HFD: 21% fat Body weight [ F1 males of HH founders (compared with F1 [52] Diet interventions (HC and HCE): males of CC founders): Adiposity (total and gonadal) Y Progressively motile sperm Y Serum cholesterol Y Sperm counts Y

(Continued on next page)

140 Trends in Endocrinology & Metabolism, February 2020, Vol. 31, No. 2 Trends in Endocrinology & Metabolism

Table 2. Continued

c c Species Treatment Effects on F0 Effects on F1 Refs

Glucose tolerance Y Sperm binding Y

Exercise intervention alone (HE) F1 males of HC founders (diet intervention

Increased serum cholesterol and insulin alone; compared with F1 males of HH resistance = founders): Interventions including exercise (HE and Percentage of progressively motile sperm in

HCE) F1 males [

Improved glucose clearance F1 males of HCE or HE founders (exercise

HE also reduced fasting glucose interventions compared with F1 males of HH All interventions: founders): Number of sperm with tail defects Y Percentage of progressively motile sperm =

Exercise interventions (HE and HCE) Sperm binding [ (only in F1 of HCE founders,

additionally restored sperm motility but not in F1 of HE founders) Diet intervention alone also increased serum All interventions in founders compared with testosterone levels HH (HFD and no intervention) or CC (control diet) founders:

Percentage of capacitated F1 sperm [

Mice: C57Bl/6 HFD: 22% fat Body weight [ F1 zygotes (recovered from normal females): [83] Abdominal and retroperitoneal adiposity [ Cleavage to the two-cell stage on day 3 and Testis weight = 4 of culture Y Plasma concentrations of cholesterol and Proportion of embryos that were on time on free fatty acids [ ﬁnal day of culture (day 5) Y Embryos at the hatching or hatched morphologic stage Y Cell count in blastocysts Y Proportion of total cells that are apoptotic [ Implantation rates of embryos Y Number of embryos that went on to form a fetus Y Fetal weight and length, placental weight, the fetal/placental weight ratio on day 18 of pregnancy =

Rats: Sprague– HFD: 22–23% fat Body weight [ F1 females: [61] Dawley Energy intake [ Body weight at day 1 Y (trend) Adiposity [ Thereafter: Plasma leptin [ Body weight = Liver mass [ Specific growth rate, energy intake, energy Skeletal muscle mass relative to body weight efficiency, adiposity, muscle mass, fasting Y plasma leptin, triglyceride, nonesterified Glucose intolerance fatty acid concentrations in adults, fasting Insulin resistance blood glucose or plasma insulin = GTT: impaired glucose tolerance Relative islet area Y (mainly owing to reduced large islets) b-Cell area Y (trend) Small islets [ (indicating a compensatory response to maintain normal b-cell mass) Altered expression of numerous genes of

(Continued on next page)

Trends in Endocrinology & Metabolism, February 2020, Vol. 31, No. 2 141 Trends in Endocrinology & Metabolism

Table 2. Continued

c c Species Treatment Effects on F0 Effects on F1 Refs

regulatory pathways associated with insulin and glucose metabolism Indications of impaired insulin-granule exocytosis mRNA expression of Il13ra2 and Ikbke [ Fos mRNA expression in islets Y Methylation at cytosine À960 of Il13ra2 in a region proximal to the transcription start site of Il13ra2 Y

Rats: Sprague– HFD: 22–23% fat F1 females: [93] Dawley Altered transcriptome in Rp-WAT Altered expression of genes related to mitochondrial and cellular responses to stress, telomerase signaling, cell death and survival, cell cycle, cellular growth and proliferation, and cancer in Rp-WAT in young females Gene expression within the olfactory transduction pathway in both Rp-WAT and pancreatic islets Y

Mice: primarily LPD: 10% protein Levels of multiple sRNAs affected in cauda Hepatic regulation of the gene encoding the [42] FVB/NJ strain sperm, including highly abundant tRFs cholesterol biosynthesis enzyme squalene background No correlation between dietary effects on epoxidase [ testicular tRNA levels and tRF changes in Ribosomal protein gene expression in LPD cauda sperm embryos Y Differences in spectrum of speciﬁc tRFs LPD embryo developmental speed Y between the testis, proximal caput epididymis, and distal cauda epididymis Consistent changes in glycine tRFs and let-7 in tissues throughout male reproductive tract, but not in the liver, muscle, or blood

Mice: C57BL/6 LPD: 9% protein Body weight, testis weight, total number of Maternal gestational weight gain, gestation [45] (for F1: sperm mature epididymal sperm = length, litter size, weight at birth = transfer or seminal Mean tubule and seminiferous epithelium Postnatal body weight of LL, NL, and LN plasma transfer) area [ offspring (compared with NN) [ Expression of the de novo and regulatory Altered levels of Enterococcus, DNA methyltransferases Dnmt1 and Dnmt3l Lactobacillus, Streptococcus,and Y Biﬁdobacterium between groups Altered expression of folate cycle enzymes Mean total bacterial content = Dhfr (dihydrofolate reductase), Mthfr NL and LN offspring: (methylenetetrahydrofolate reductase), and Serum TNF Y and hepatic expression of the Mtr (methionine synthase) in testis central NAFLD-regulating genes [ Global level of sperm-DNA methylation Y All offspring derived from LPD sperm and/ or seminal plasma developed phenotypic characteristics of metabolic dysfunction

(Continued on next page)

142 Trends in Endocrinology & Metabolism, February 2020, Vol. 31, No. 2 Trends in Endocrinology & Metabolism

Table 2. Continued

c c c Species Treatment Effects on F0 Effects on F1 Effects on F2 Refs

Rats: HFD: 21– Body weight [ Body weight at 3 days of age Y Birth-weight Y [17]

Sprague– 23% Impaired glucose b-Cell mass at 5 days of age Y GTT: glucose AUC in F2 fed a HFD

Dawley tolerance GTT: glucose AUC in F1 fed HFD (compared with F2 fed chow) [

Several differentially (compared with F1 fed chow) [ F2 females resistant to HFD-induced expressed miRNAs, Females resistant to HFD-induced weight gain

piRNAs and tRFs weight gain F2 males fed chow during GTT: insulin Several differentially expressed sncRNAs levels [ Changes in let-7c expression

Numerous DMR in the spermatozoa of F0 and F1 fed a HFD

Altered expression of several tRFs and miRNA in both F0 fathers

and their F1 sons only when fed a HFD Differential expression of three piRNAs and let-7c miRNA in the

sperm of F0 fathers and their F1 sons

Mice: HFD: 22% Total body weight [ F1 males: F1 females: F2 males: F2 females: [31]

C57BL/6 fat ROS levels [ Testes and seminal Meiotic F1 fathers: F1 fathers: Sperm DNA vesicle weights, competence of Reproductive Oocyte damage [ testosterone oocytes Y organ weights = oxidative stress Percent of levels = Altered Testosterone [ morphologically Sperm function mitochondrial levels = Altered abnormal sperm [ including motility Y membrane Sperm motility Y mitochondrial Glucose Sperm ROS and potential in all ROS levels [ function

homeostasis by DNA damage [ regions of the F1 mothers: F1 mothers: fasting blood Binding to oocyte oocytes Testes and Ovary weights = glucose, GTT, or Y (marker of seminal vesicles Total number of plasma insulin Fertilization Y subfertility) (as a proportion oocytes = levels = Mating behavior, of total body Cumulus cell Successful matings time to mate, weight) Y expansion with females Y gestational length, Testosterone index = Time to mate, number of pups levels Y Meiotic

gestational length, sired by F1 males = Sperm motility Y progression of litter size = ROS levels [ oocytes =

Mice: HFD: 21% Body weight [ F1 males: F1 females: F2 males: F2 females: [2]

C57BL/6 fat Body fat [ Impaired glucose Body weight F1 fathers: F1 fathers: Glucose tolerance from from day 5 into No adiposity Insulin homeostasis = 8 weeks of age adulthood Total body resistance [ Fasting insulin onward, persisting despite control weight, body Adiposity [ levels = up to the age of diet [ composition = Body weight = Insulin sensitivity = 26 weeks Obesity [ Plasma leptin Y Summed Sperm ROS levels [ Insulin resistance Adipose tissue No altered adipose depots Sperm DNA evident by depots in both glucose as a proportion damage [ 26 weeks of age absolute terms tolerance or of body weight 11 Differentially and as a insulin sensitivity [

expressed miRNAs proportion of F1 mothers: Liver and in testes body weight [ Body weight [ pancreas weight 4 miRNAs Circulating lipids Adipose depots [ dysregulated in [ in absolute Impaired insulin

(Continued on next page)

Trends in Endocrinology & Metabolism, February 2020, Vol. 31, No. 2 143 Trends in Endocrinology & Metabolism

Table 2. Continued

c c c Species Treatment Effects on F0 Effects on F1 Effects on F2 Refs

sperm terms [ sensitivity

qPCR conﬁrmed Impaired F1 mothers: dysregulation of 4 glucose Total body mRNAs (Glt28d2, tolerance and weight = Hmgb1, Lrrtm3, and insulin resistance Impaired insulin Smarce1) in HFD sensitivity testis (all of these are predicted to be commonly targeted by at least 2 of the 4 differentially expressed miRNAs in testes and sperm) Level of global DNA methylation in testes and late elongated spermatids Y aTreatments include a high-fat diet (HFD) or low-protein diet (LPD) in rats or mice. bKey and abbreviations: Key, [, increase; Y, decrease; =, no change; AUC, area under the curve; CC, baseline control: CD during the initial feeding period and CD during the intervention period; CD, control diet; DMR, differentially methylated region; EPM, elevated plus maze; GTT, glucose tolerance test; HC, change to a CD; HCE; change to a CD with exercise; HE, HFD continuation with exercise; HH, HFD continuation; hsp90, heat shock protein 90; let-7c, regulates the expression of metabolic genes in adipose tissue; LG-ABN, licking/grooming and arched-back nursing; LL, LPD sperm and LPD seminal plasma; LN, LPD sperm and NPD seminal plasma; NAFLD, non-alcoholic fatty liver disease; NL, NPD sperm and LPD seminal plasma; NN, NPD sperm and NPD seminal plasma; NPD, normal protein diet; OF, open ﬁeld; ROS, reactive oxygen species; Rp-WAT, retroperitoneal white adipose tissue; SFA, short-chain fatty acids; tRF, tRNA fragments. cCompared with the control group.

A substantial body of evidence links obesity with increased cortisol activity, indicative of alterations in the hypothalamic–pituitary–adrenal (HPA) axis [90]. Male F0 generation mice fed an obesogenic diet have higher basal corticosterone levels as well as increased HPA axis responsiveness to restraint stress than control-fed conspeciﬁcs [91]. However, whether obesity contributes to HPA axis dysregulation or vice versa is currently unclear. In adult rat offspring, long-term effects of a maternal HFD were indeed observed on basal HPA axis activity, withelevatedmeanplasmacorticosteronelevels[92]. Future studies will be necessary to clarify whether paternal dietary patterns can also affect the HPA axis of the offspring.

In further molecular studies, female mice sired by HFD-fed males were found to display changes in the expression of genes involved in the pathogenesis of obesity and diabetes, such as mitogen-activated protein kinase (MAPK) pathway genes [93]. Additional effects of paternal HFD included histopathological signs of non-alcoholic fatty liver (hepatic steatosis) and reduced liver function in offspring [24,94]. Notably, an association has been revealed between fatty liver disease in humans, gut dysbiosis, and a shift in gut metabolic function [95]. Whether the altered liver function is a result of paternal diet inﬂuences on the altered intestinal microbiota of the offspring remains unclear.

Recently, metabolism and immunity have been found to be much more closely interrelated than was previously thought [96]. Saturated fatty acids, which are typical for Western diets, appear to be an important risk factor for prompting a proinflammatory response in immune cells. Furthermore, maternal obesity has been linked to inflammation and immune dysregulation in newborns [97]. The role of paternal obesity and the long-term consequences for the offspring as well as the underlying mechanisms are yet to be defined.

144 Trends in Endocrinology & Metabolism, February 2020, Vol. 31, No. 2 Trends in Endocrinology & Metabolism

Cardiovascular Disease and Cancer Risk In recent years, increasing evidence has emerged that parental obesity increases the susceptibility to cardiovascular disease [98] and cancer development [29,99,100] in the offspring. Both maternal and paternal BMI were found to be negatively associated with offspring cardiorespiratory ﬁtness, even in the absence of overweight in the progeny [98]. As mentioned before, paternal obesity- induced IGF2 imprinting defects have been associated with carcinogenesis. Furthermore, a paternal history of obesity has been linked to an accelerated development of breast cancer [29] or pancreatic cancer [99] in the respective animal models (i.e., a rat model of mammary carcinogenesis and a mouse model of pancreatic cancer). Higher rates of mammary cancer were also observed in daughters of male mice exposed to a low-protein diet [100]. Hence, paternal poor-quality diets seem to play a role in the programming of offspring cancer risk, thereby supporting the concept of the epigenetic inheritance of cancer risk. However, the detailed mechanisms, and the full extent of preconceptual environmental exposures and offspring cardiovascular and cancer susceptibilities, remain elusive.

Paternal Diet and Obesity-Evoked Inﬂuences on Offspring Brain and Behavior Despite the growing body of knowledge of paternal dietary effects on the physiology of the offspring, less is known about changes in their behavior. The complexity of the relationship is made evident by inconsistent intergenerational effects of maternal HFD and/or obesity, including alterations to anxiety- and depressive-like behaviors, cognitive development, and locomotor activity/exploration in the offspring (e.g., [101,102]). Overall, mounting evidence links maternal overnutrition with cognitive def- icits (reviewed in [103]) and neurobiological alterations to neuronal morphology in the hippocampus and amygdala in the offspring [104]. Furthermore, the offspring of male HFD-fed mice displayed impaired learning and memory, although they were never exposed to the HFD themselves [105]. As such, the question is raised as to what potential neurobiological mechanisms might mediate paternal effects on offspring higher-order cognition.

Counteractive Effects of Exercise and Healthy Diet As opposed to a ‘junk food’ diet and a sedentary lifestyle, paternal caloric restriction and aerobic exercise before conception are associated with improved cognition and reduced anxiety-like behavior in the progeny in rodents [19,106,107]. From an evolutionary perspective, this intergenerational programming effect may be adaptive and prepare the offspring in the event of food scarcity. Improved spatial memory, sociality, and decreased anxiety, for example, could help with food acquisition as well as with reproduction. However, in many modern societies, a major ecological factor driving the evolution of advanced cognitive skills, namely food scarcity, is not present [108]. Furthermore, little is known about whether adaptive processes can occur via paternal epigenetic inheritance [109]. The transmission of obesity-induced changes in emotional behavior to the next generation could create a vulnerability for the subsequent development of mood disorders in the offspring [110,111]. However, the transgenerational effects of diet and obesity may only become apparent when combined with further stressors, such as traumatic events, chronic stress, or further consumption of poor diet. Indeed, evidence suggests that a maternal HFD increases the sensitivity of the offspring to the proinﬂamma- tory effects of stress or an unhealthy diet [112]. By contrast, a healthy diet and exercise during the perinatal period may be beneﬁcial in the prevention of mood disorders [19,106,107].

Concluding Remarks and Future Perspectives Most studies to date have focused on the impact of either high-fat or high-sugar diets on metabolic, physiologic, and psychiatric parameters. However, disentangling the effects of single macronutrients to study the relationship between diet and disease is not currently feasible, which is why a comprehensive analysis of the combined effects of HFHS diets would be of greater signiﬁcance. It has been shown that other forms of suboptimal paternal nutrition can have deleterious effects on the offspring. For example, investigations are needed on whether plant-based proteins have a different effect than animal-based proteins, as has been demonstrated for plant- and animal-based fats (see Outstanding Questions) [29].

Trends in Endocrinology & Metabolism, February 2020, Vol. 31, No. 2 145 Trends in Endocrinology & Metabolism

Some studies found significant diet-induced effects on various health parameters based on feeding Outstanding Questions so-called ‘cafeteria diets’. However, the results seem to be of limited significance because the IVF commonly generates embryos observed effects cannot be traced back to single macronutrients but could be caused by specific in- that develop in a seminal plasma- gredients such as artificial additives or preservatives. Likewise, the use of homogenized pelleted diets free environment. The fact that composed of >60% fat (e.g., [23,29,79]) most likely leads to a lack of important nutrients and soluble/ recent studies have demonstrated insoluble fiber, possibly resulting in deleterious effects that are not only the result of a high-calorie that seminal plasma exerts signifi- diet. As such, dietary specificity and well-matched controls are crucial in the study of dietary impacts cant effects on the developing em- on obesity to allow the molecular and cellular mechanisms to be systematically investigated. bryo challenges prevailing IVF Commercially available rodent foods that mimic a human Western diet contain proportionally techniques and raises the question high-fat and high-sugar contents, and allow the effects of HFHS diets to be studied in a controlled, of whether IVF could impact on standardized way ([113] for a comparison of different HFDs). the health status and disease susceptibility of offspring.

Furthermore, evidence points towards a crucial role of the gut microbiota on metabolism and its Ultra-processed foods evoke high impact on brain and behavior via the gut–brain axis, and also of maternal ‘seeding’ of the microbiota levels of salt consumption; howev- in her offspring [114,115]. The paternal inﬂuence on the microbiota is more complex, particularly er, the transgenerational impact of because dams might take up gut bacteria of HFD-fed males through coprophagy at time of mating. excessivesaltintakeonphysiology This behavior might cause a change in the microbiome of the dam, which could in turn affect the mi- and behavior remains to be ad- crobiota of the offspring and eventually their behavior. Therefore, there is some evidence that the dressed. How do ultra-processed foods and increased salt intake programming effect of a paternal HFD on offspring [2] might actually be transmitted via the paternal, impact on reproductive function and then the by maternal, gut microbiome. and fertility in males and females?

Childhood obesity has emerged as a global health epidemic and is a major public health issue owing Do highly palatable ‘cafeteria diets’ to the associated comorbidities of obesity, underlining the importance of developing prevention or homogenized high-fat or high- methods. However, the development of preventive techniques requires a comprehensive under- sugar pelleted diets in rodents adequately represent a ‘Western standing of all contributing factors. For a long time, paternal nongenetic effects on offspring pheno- diet’ in humans? What conclusions types have been neglected and many questions remain. Future research will be necessary to identify canbedrawnonthebasisofa the underlying mechanisms and relatedness between the different molecular processes involved. A diet replete with artificial additives comprehensive knowledge of the role of paternally driven epigenetic inheritance of dietary patterns and preservatives that also most and obesity may assist in the development of prevention and intervention techniques to improve the likely causes a deficiency of impor- health of future generations. tant nutrients and soluble/insoluble fiber?

Acknowledgments Do diets enriched in monosatu- A.C.R. receives funding from a Canada First Research Excellent Fund BrainsCAN Senior Fellowship. rated fatty acids, u-3 polyunsatu- A.J.H. is a National Health and Medical Research Council (NHMRC) Principal Research Fellow and is rated fatty acids, or vitamins have also supported by NHMRC Project Grants and the DHB Foundation Equity Trustees. This work was enhancing transgenerational ef- funded by an Australian Research Council (ARC) Discovery Project grant (DP180101974 to A.C.R. fects, and can they reverse the adverse effects of a paternal poor- and A.J.H.). The Florey Institute of Neuroscience and Mental Health acknowledges the strong support quality saturated fatty acid-rich from the Victorian Government and in particular the funding from the Operational Infrastructure Sup- HFD in the offspring? port Grant. Does paternal obesity affect the reproductive health of children References and grandchildren in humans? 1. Chambers, T.J.G. et al. (2016) High-fat diet disrupts 5. Zhang, T. et al. (2019) Rate of change in body mass metabolism in two generations of rats in a parent- index at different ages during childhood and adult What are the transgenerational im- of-origin speciﬁc manner. Sci. Rep. 6, 31857 obesity risk. Pediatr. Obes. 14, e12513 pacts of diets high in plant-based 2. Fullston, T. et al. (2013) Paternal obesity initiates 6. Whitaker, R.C. et al. (1997) Predicting obesity in proteins versus animal-based pro- metabolic disturbances in two generations of mice young adulthood from childhood and parental with incomplete penetrance to the F2 generation obesity. N. Engl. J. Med. 337, 869–873 teins? and alters the transcriptional proﬁle of testis and 7. Castellini, G. et al. (2017) Fat mass and obesity- sperm microRNA content. FASEB J. 27, 4226–4243 associated gene (FTO) is associated to eating 3. National Center for Health Statistics (1999–2016) disorders susceptibility and moderates the National Health and Nutrition Examination Survey expression of psychopathological traits. PLoS One (NHANES), Centers for Disease Control and 12, e0173560 Prevention (CDC). www.cdc.gov/nchs/data/ 8. Nilsson, E.E. et al. (2018) Environmentally induced nhanes/index.htm epigenetic transgenerational inheritance of 4. World Health Organization. (2018) Obesity and disease. Environ. Epigenet. 4, dvy016 Overweight, WHO. www.who.int/news-room/fact- 9. Bird, A. (2007) Perceptions of epigenetics. Nature sheets/detail/obesity-and-overweight 447, 396–398

146 Trends in Endocrinology & Metabolism, February 2020, Vol. 31, No. 2 Trends in Endocrinology & Metabolism

10. Skvortsova, K. et al. (2018) Functions and 30. Ferramosca, A. et al. (2017) Dietary fatty acids mechanisms of epigenetic inheritance in animals. influence sperm quality and function. Andrology 5, Nat. Rev. Mol. Cell Biol. 19, 774–790 423–430 11. Hales, C.M. et al. (2018) Trends in obesity and 31. Fullston, T. et al. (2012) Diet-induced paternal severe obesity prevalence in US youth and adults by obesity in the absence of diabetes diminishes the sex and age, 2007–2008 to 2015–2016. JAMA 319, reproductive health of two subsequent generations 1723–1725 of mice. Hum. Reprod. 27, 1391–1400 12. Soubry, A. (2018) Epigenetics as a driver of 32. Ferigolo, P.C. et al. (2019) Sperm functional aspects developmental origins of health and disease: did and enriched proteomic pathways of seminal we forget the fathers? Bioessays 40, 1700113 plasma of adult men with obesity. Andrology 7, 13. Korgan, A.C. et al. (2018) Effects of paternal high-fat 341–349 diet and rearing environment on maternal 33. Javurek, A.B. et al. (2016) Discovery of a novel investment and development of defensive seminal fluid microbiome and influence of estrogen responses in the offspring. receptor alpha genetic status. Sci. Rep. 6, 23027 Psychoneuroendocrinology 91, 20–30 34. Javurek, A.B. et al. (2017) Consumption of a high- 14. Ornellas, F. et al. (2017) Obese fathers lead to an fat diet alters the seminal fluid and gut altered metabolism and obesity in their children in microbiomes in male mice. Reprod. Fertil. Dev. adulthood: review of experimental and human 29, 1602–1612 studies. J. Pediatr. (Rio J) 93, 551–559 35. Belva, F. et al. (2012) Are ICSI adolescents at risk 15. Fleming, T.P. et al. (2018) Origins of lifetime health for increased adiposity? Hum. Reprod. 27, around the time of conception: causes and 257–264 consequences. Lancet 391, 1842–1852 36. Belva, F. et al. (2016) Semen quality of young adult 16. Velazquez, M.A. et al. (2019) Periconceptional ICSI offspring: the first results. Hum. Reprod. 31, environment and the developmental origins of 2811–2820 disease. J. Endocrinol. 242, 33–49 37. Rodrı´guez-Martıńez, H. et al. (2011) Seminal plasma 17. Barbosa, Castro de, T. et al. (2016) High-fat diet proteins: what role do they play? Am. J. Reprod. reprograms the epigenome of rat spermatozoa and Immunol. 66 (Suppl. 1 ), 11–22 transgenerationally affects metabolism of the 38. Robertson, S.A. (2005) Seminal plasma and male offspring. Mol. Metab. 5, 184–197 factor signalling in the female reproductive tract. 18. Grandjean, V. et al. (2015) RNA-mediated paternal Cell Tissue Res. 322, 43–52 heredity of diet-induced obesity and metabolic 39. Sharkey, D.J. et al. (2012) Seminal fluid induces disorders. Sci. Rep. 5, 18193 leukocyte recruitment and cytokine and chemokine 19. Yeshurun, S. and Hannan, A.J. (2019) mRNA expression in the human cervix after coitus. Transgenerational epigenetic influences of paternal J. Immunol. 188, 2445–2454 environmental exposures on brain function and 40. Samanta, L. et al. (2018) The enigmatic seminal predisposition to psychiatric disorders. Mol. plasma: a proteomics insight from ejaculation to Psychiatry 24, 536–548 fertilization. Reprod. Biol. Endocrinol. 16, 41 20. Pang, T.Y. et al. (2017) Transgenerational paternal 41. Reilly, J.N. et al. (2016) Characterisation of mouse transmission of acquired traits: Stress-induced epididymosomes reveals a complex profile of modification of the sperm regulatory transcriptome microRNAs and a potential mechanism for and offspring phenotypes. Curr. Opin. Behav. Sci. modification of the sperm epigenome. Sci. Rep. 6, 14, 140–147 31794 21. Bakos, H.W. et al. (2011) The effect of paternal diet- 42. Sharma, U. et al. (2016) Biogenesis and function of induced obesity on sperm function and fertilization tRNA fragments during sperm maturation and in a mouse model. Int. J. Androl. 34, 402–410 fertilization in mammals. Science 351, 391 22. Fullston, T. et al. (2017) The most common vices of 43. Bromfield, J.J. et al. (2014) Maternal tract factors men can damage fertility and the health of the next contribute to paternal seminal fluid impact on generation. Journal of Endocrinology 234, F1–F6 metabolic phenotype in offspring. PNAS 111, 2200– 23. Chen, Q. et al. (2016) Sperm tsRNAs contribute to 2205 intergenerational inheritance of an acquired 44. Saccone, G. et al. (2019) Effectiveness of seminal metabolic disorder. Science 351, 397–400 plasma in in vitro fertilisation treatment: a 24. Fan, Y. et al. (2015) Diet-induced obesity in male systematic review and meta-analysis. BJOG 126, C57BL/6 mice decreases fertility as a consequence 220–225 of disrupted blood–testis barrier. PLoS ONE 10, 45. Watkins, A.J. et al. (2018) Paternal diet programs e0120775 offspring health through sperm- and seminal 25. Martini, A.C. et al. (2010) Overweight and seminal plasma-specific pathways in mice. PNAS 115, quality: a study of 794 patients. Fertil. Steril. 94, 10064–10069 1739–1743 46. Hammoud, S.S. et al. (2009) Distinctive chromatin in 26. Chavarro, J.E. et al. (2010) Body mass index in human sperm packages genes for embryo relation to semen quality, sperm DNA integrity, and development. Nature 460, 473–478 serum reproductive hormone levels among men 47. Schagdarsurengin, U. and Steger, K. (2016) attending an infertility clinic. Fertil. Steril. 93, 2222– Epigenetics in male reproduction: effect of paternal 2231 diet on sperm quality and offspring health. Nat. Rev. 27. Crean, A.J. and Senior, A.M. (2019) High-fat diets Urol. 13, 584–595 reduce male reproductive success in animal 48. Carone, B.R. et al. (2010) Paternally induced models: a systematic review and meta-analysis. transgenerational environmental reprogramming Obes. Rev. 20, 921–933 of metabolic gene expression in mammals. Cell 143, 28. Ferramosca, A. et al. (2016) A high-fat diet 1084–1096 negatively affects rat sperm mitochondrial 49. Yamaguchi, K. et al. (2018) Re-evaluating the respiration. Andrology 4, 520–525 localization of sperm-retained histones revealed the 29. Fontelles, C.C. et al. (2016) Paternal programming modification-dependent accumulation in specific of breast cancer risk in daughters in a rat model: genome regions. Cell Rep. 23, 3920–3932 opposing effects of animal- and plant-based high- 50. Erkek, S. et al. (2013) Molecular determinants of fat diets. Breast Cancer Res. 18, 71 nucleosome retention at CpG-rich sequences in

Trends in Endocrinology & Metabolism, February 2020, Vol. 31, No. 2 147 Trends in Endocrinology & Metabolism

mouse spermatozoa. Nat. Struct. Mol. Biol. 20, 72. Nixon, B. et al. (2015) The microRNA signature of 868–875 mouse spermatozoa is substantially modified 51. Saitou, M. and Kurimoto, K. (2014) Paternal during epididymal maturation. Biol. Reprod. 93, 91 nucleosomes: are they retained in developmental 73. Ostermeier, G.C. et al. (2004) Reproductive biology: promoters or gene deserts? Dev. Cell 30, 6–8 delivering spermatozoan RNA to the oocyte. 52. McPherson, N.O. et al. (2014) Obese father’s Nature 154 metabolic state, adiposity, and reproductive 74. Gaskins, A.J. et al. (2012) Dietary patterns and capacity indicate son’s reproductive health. Fertil. semen quality in young men. Hum. Reprod. 27, Steril. 101, 865–873 2899–2907 53. Zhang, Y. et al. (2019) Sperm RNA code 75. Daxinger, L. and Whitelaw, E. (2012) programmes the metabolic health of offspring. Nat. Understanding transgenerational epigenetic Rev. Endocrinol. 15, 489–498 inheritance via the gametes in mammals. Nat. Rev. 54. Sharma, U. et al. (2018) Small RNAs are trafficked Genet. 13, 153–162 from the epididymis to developing mammalian 76. Morgan, C.P. et al. (2019) Driving the next sperm. Dev. Cell 46, 481–494 generation: paternal lifetime experiences 55. Schaefer, C.B. et al. (2007) Epigenetic decisions in transmitted via extracellular vesicles and their small mammalian germ cells. Science 316, 398 RNA cargo. Biol. Psychiatry 85, 164–171 56. Kaneda, M. et al. (2004) Essential role for de novo 77. Cropley, J.E. et al. (2016) Male-lineage transmission DNA methyltransferase Dnmt3a in paternal and of an acquired metabolic phenotype induced by maternal imprinting. Nature 429, 900–903 grand-paternal obesity. Mol. Metab. 5, 699–708 57. Tunc, O. and Tremellen, K. (2009) Oxidative DNA 78. Schorn, A.J. et al. (2017) LTR-retrotransposon damage impairs global sperm DNA methylation in control by tRNA-derived small RNAs. Cell 170, infertile men. J. Assist. Reprod. Genet. 26, 537–544 61–71 58. Lee, H.J. et al. (2014) Reprogramming the 79. An, T. et al. (2017) Long non-coding RNAs could act methylome: erasing memory and creating diversity. as vectors for paternal heredity of high fat diet- Cell Stem Cell 14, 710–719 induced obesity. Oncotarget 8, 47876–47889 59. Kobayashi, H. et al. (2012) Contribution of intragenic 80. Gapp, K. et al. (2018) Alterations in sperm long RNA DNA methylation in mouse gametic DNA contribute to the epigenetic inheritance of the methylomes to establish oocyte-specific heritable effects of postnatal trauma. Mol. Psychiatry. marks. PLoS Genet 8, e1002440 Published online October 30, 2018. https://doi.org/ 60. Radford, E.J. et al. (2014) In utero undernourishment 10.1038/s41380-018-0271-6. perturbs the adult sperm methylome and 81. Morris, K.V. and Mattick, J.S. (2014) The rise of intergenerational metabolism. Science 345, regulatory RNA. Nat. Rev. Genet. 15, 423–437 1255903 82. Luo, Z. et al. (2019) Comprehensive analysis of 61. Ng, S.-F. et al. (2010) Chronic high-fat diet in fathers differences of N6-methyladenosine RNA programs b-cell dysfunction in female rat offspring. methylomes between high-fat-fed and normal Nature 467, 963–966 mouse livers. Epigenomics 11, 1267–1282 62. de Waal, E. et al. (2012) Primary epimutations 83. Mitchell, M. et al. (2011) Paternal diet-induced introduced during intracytoplasmic sperm injection obesity impairs embryo development and (ICSI) are corrected by germline-specific epigenetic implantation in the mouse. Fertil. Steril. 95, 1349– reprogramming. Proc. Natl. Acad. Sci. U. S. A. 109, 1353 4163–4168 84. Raad, G. et al. (2019) Adverse effects of paternal 63. Skinner, M.K. et al. (2018) Alterations in sperm DNA obesity on the motile spermatozoa quality. PLoS methylation, non-coding RNA and histone ONE 14, e0211837 retention associate with DDT-induced epigenetic 85. Luna, M. et al. (2009) Paternal age and assisted transgenerational inheritance of disease. reproductive technology outcome in ovum Epigenetics Chromatin 11, 8 recipients. Fertil. Steril. 92, 1772–1775 64. Ben Maamar, M. et al. (2019) Developmental origins 86. Kort, H.I. et al. (2006) Impact of body mass index of transgenerational sperm DNA methylation values on sperm quantity and quality. J. Androl. 27, epimutations following ancestral DDT exposure. 450–452 Dev. Biol. 445, 280–293 87. Soubry, A. et al. (2013) Paternal obesity is associated 65. Jodar, M. et al. (2013) The presence, role and clinical with IGF2 hypomethylation in newborns: results use of spermatozoal RNAs. Hum. Reprod. Update from a Newborn Epigenetics Study (NEST) cohort. 19, 604–624 BMC Med. 11, 29 66. Chen, Q. et al. (2016) Epigenetic inheritance of 88. Power, C. et al. (2003) Combination of low acquired traits through sperm RNAs and sperm birth weight and high adult body mass index: at RNA modifications. Nat. Rev. Genet. 17, 733–743 what age is it established and what are its 67. Rodgers, A.B. et al. (2015) Transgenerational determinants? J. Epidemiol. Community Health 57, epigenetic programming via sperm microRNA 969–973 recapitulates effects of paternal stress. PNAS 112, 89. Consitt, L.A. et al. (2018) Paternal high-fat diet 13699–13704 enhances offspring whole-body insulin sensitivity 68. Donkin, I. et al. (2016) Obesity and bariatric surgery and skeletal muscle insulin signaling early in life. drive epigenetic variation of spermatozoa in Physiol. Rep. 6, e13583 humans. Cell Metab. 23, 369–378 90. Incollingo Rodriguez, A.C. et al. (2015) 69. Gapp, K. et al. (2014) Implication of sperm RNAs in Hypothalamic–pituitary–adrenal axis dysregulation transgenerational inheritance of the effects of early and cortisol activity in obesity: a systematic review. trauma in mice. Nat. Neurosci. 17, 667–669 Psychoneuroendocrinology 62, 301–318 70. Zhang, Y. et al. (2018) Dnmt2 mediates 91. Sharma, S. and Fulton, S. (2013) Diet-induced intergenerational transmission of paternally obesity promotes depressive-like behaviour that is acquired metabolic disorders through sperm small associated with neural adaptations in brain reward non-coding RNAs. Nat. Cell Biol. 20, 535–540 circuitry. Int. J. Obes. (Lond) 37, 382–389 71. Conine, C.C. et al. (2018) Small RNAs gained during 92. Niu, X. et al. (2019) Maternal high fat diet programs epididymal transit of sperm are essential for hypothalamic-pituitary-adrenal function in adult embryonic development in mice. Dev. Cell 46, rat offspring. Psychoneuroendocrinology 102, 470–480 128–138

148 Trends in Endocrinology & Metabolism, February 2020, Vol. 31, No. 2 Trends in Endocrinology & Metabolism

93. Ng, S.-F. et al. (2014) Paternal high-fat diet remodeling in adult male offspring. consumption induces common changes in the Psychoneuroendocrinology 83, 49–57 transcriptomes of retroperitoneal adipose and 105. Zhou, Y. et al. (2018) Diet-induced paternal obesity pancreatic islet tissues in female rat offspring. impairs cognitive function in offspring by mediating FASEB J. 28, 1830–1841 epigenetic modifications in spermatozoa. Obesity 94. Zhang, X. et al. (2019) Paternal programming of liver 26, 1749–1757 function and lipid profile induced by a paternal pre- 106. Govic, A. et al. (2016) Paternal calorie restriction conceptional unhealthy diet: potential association prior to conception alters anxiety-like behavior of with altered gut microbiome composition. Kidney the adult rat progeny. Psychoneuroendocrinology Blood Press. Res. 44, 133–148 64, 1–11 95. Boursier, J. et al. (2016) The severity of nonalcoholic 107. Short, A.K. et al. (2017) Exercise alters mouse sperm fatty liver disease is associated with gut dysbiosis small noncoding RNAs and induces a and shift in the metabolic function of the gut transgenerational modification of male offspring microbiota. Hepatology 63, 764–775 conditioned fear and anxiety. Transl. Psychiatry 7, 96. Billi, A.C. and Gudjonsson, J.E. (2019) Overriding e1114 the immune system’s sweet tooth: fatty acids rile up 108. Mattson, M.P. (2019) An evolutionary perspective on innate immunity. Cell 177, 1088–1090 why food overconsumption impairs cognition. 97. Broadney, M.M. et al. (2017) Impact of parental Trends Cogn. Sci. 23, 200–212 obesity on neonatal markers of inflammation and 109. Fullston, T. et al. (2015) Paternal obesity induces immune response. Int. J. Obes. (Lond) 41, 30–37 metabolic and sperm disturbances in male offspring 98. Labayen, I. et al. (2010) Intergenerational that are exacerbated by their exposure to an cardiovascular disease risk factors involve both ’obesogenic’ diet. Physiol. Rep. 3, e12336 maternal and paternal BMI. Diabetes Care 33, 110. Sasaki, A. et al. (2013) Perinatal high fat diet alters 894–900 glucocorticoid signaling and anxiety behavior in 99. da Cruz, R.S. et al. (2019) Parental obesity programs adulthood. Neuroscience 240, 1–12 pancreatic cancer development in offspring. 111. Bilbo, S.D. and Tsang, V. (2010) Enduring Endocr. Relat. Cancer. .Published online March 1, consequences of maternal obesity for brain 2019. https://doi.org/10.1530/ERC-19-0016. inflammation and behavior of offspring. FASEB J. 100. da Cruz, R.S. et al. (2018) Paternal malnutrition 24, 2104–2115 programs breast cancer risk and tumor metabolism 112. Lin, C. et al. (2015) Maternal high fat diet programs in offspring. Breast Cancer Res 20, 99 stress-induced behavioral disorder in adult 101. Sullivan, E.L. et al. (2015) Maternal high-fat diet offspring. Physiol. Behav. 152, 119–127 programming of the neuroendocrine system and 113. Bortolin, R.C. et al. (2018) A new animal diet based behavior. Horm. Behav. 76, 153–161 on human Western diet is a robust diet-induced 102. Zieba, J. et al. (2019) Long-term behavioural effects obesity model: comparison to high-fat and of maternal obesity in C57BL/6J mice. Physiol. cafeteria diets in term of metabolic and gut Behav. 199, 306–313 microbiota disruption. Int. J. Obes. (Lond) 42, 103. Contu, L. and Hawkes, C.A. (2017) A review of the 525–534 impact of maternal obesity on the cognitive function 114. Powell, C.M. (2017) Mum’s bacteria linked to baby’s and mental health of the offspring. Int. J. Mol. Sci. behaviour. Nature, 466–467 18, E1093 115. Kim, S. et al. (2017) Maternal gut bacteria promote 104. Janthakhin, Y. et al. (2017) Maternal high-fat diet neurodevelopmental abnormalities in mouse leads to hippocampal and amygdala dendritic offspring. Nature 549, 528–532

Trends in Endocrinology & Metabolism, February 2020, Vol. 31, No. 2 149