Science in Society Archive

New Genetics, Evolution, & Hazards of GMOs

Sperm-Mediated Inheritance of Acquired Characters

Heredity does not just depend on DNA in chromosomes or organelles, loose sequences of RNA and DNA can be independently inherited, and readily taken up by sperm cells to transfer the father’s acquired characters to the fertilized egg; this underscores the trans-generational hazards of new toxins and nucleic acids introduced into our food chain via genetically modified organisms Dr. Mae-Wan Ho

Fathers are good at passing on acquired characters

The first hint that fathers can pass on acquired characters was the discovery that the experience of young boys could affect not just their health in later life, but also the health of their sons and grandsons. That was the beginning of the epigenetic revolution [1] (Epigenetic Inheritance - What Genes Remember, SiS 41). All kinds of life experiences, good and bad, from caring mothers to environmental toxins, leave epigenetic imprints that are passed on for generations afterwards (see [2, 3] Caring Mothers Strike Fatal Blow against Genetic Determinism, and Epigenetic Toxicology, SiS 41). In the case of environmental toxins, Michael Skinner’s reproductive biology lab at Washington State University Pullman in the United States first reported in 2005 that injecting pregnant rats with endocrine disruptor fungicide vinclozolin caused sperm abnormalities that persisted in the male progeny for at least 4 generations [4]. The effects on reproduction correlate with altered DNA methylation pattern in the germ line (though the methylation differences vary widely among the animals, and failed to satisfy his critics [5]). Subsequently, they found that insecticides DDT and permethrin, jet fuel, plastic additives phthalates and bisphenol A, and dioxin can all trigger trans-generational health effects in rats such as obesity and ovarian disease, and each resulted in a different pattern of methylation in sperm DNA, according to Skinner.

DNA methylation is not the only means of transmitting acquired epigenetic information to subsequent generations, chromatin modification (via histone and RNA) as well as various non-coding (nc) RNAs are also involved (reviewed in [6]). NcRNAs such as those involved in RNA interference can be independently inherited, and can also direct chromatin modification and DNA methylation see [7, 8] RNA Inheritance of Acquired Characters, and Nucleic Acid Invaders from Food Confirmed, SiS 63).

Cytosine methylation of DNA involves a subset of genomic cytosines methylated at the C5 position in some species. In mammals, it occurs in CpG dinucleotides, whereas in plants non-CpG cytosines can also be methylated [6]. It is implicated in many of the best established epigenetic inheritance paradigm; though major model organisms such as worms and flies have perfectly functional epigenetic inheritance despite lacking cytosine methylation. Soon after fertilization, the vast majority of methylcytosine in sperm is converted by the Tet3 enzyme to hydroxymethylcytosine that gets lost by dilution during replication, effectively erasing cytosine methylation patterns except for a subset that is maintained, including those of some imprinted genes. Conversely, maternal cytosine methylation is protected from hydroxylation by the PGC7/Dppa3/Stella protein and methylation pattern is effectively maintained. PGC7/Dpp3a/Stella is targeted to the genome via binding to the heterochromatic histone mark H3K9me2 (this refers to a specific amino acid lysine in position 9 on histone H3 that has 2 methyl groups added). H3K9me2 was found at several paternally methylated imprinted regions in sperm, suggesting that this histone mark signals special sites of the paternal genome where methylation is maintained.

Eukaryote genomes are packaged into a nucleoprotein complex known as chromatin. Germ cell chromatin is vastly different from that of other cells. In mammals, most histone proteins are lost during sperm development, eventually to be replaced by protamines. However, genes expressed early in development may preferentially retain histones. After fertilization, the sperm genome is rapidly stripped of protamines and most, but not all histones.  

Germ cells of many organisms carry many RNA species that can affect the phenotype of offspring, and can be paternally inherited across generations (see [7]).

What complicates matters is that every known epigenetic information carrier interacts with every other carrier. Cytosine modifications directly affect the positioning of nucleosomes (the first level of chromatin organization into chromosome involving the DNA chain wound around a core of 8 histone proteins), and recruit chromatin-modifying complexes that modify histones. Conversely histone modifications can affect recruitment of cytosine methylases and demethylases. Small RNAs including short-interfering (si)RNAs and piRNAs, and long RNAs such as long intergenic noncoding (linc)RNAs can direct histone modifications and cytosine methylation. And to complete the circle, chromatin structure and DNA modifications affect transcription of small RNA and lncRNA-containing loci. Such circular cross talk will complicate analysis of epigenetic marks carried out in many studies, because they may be finding downstream effects of some original and perhaps long-erased epigenetic perturbation (as in the example described below).

Epigenetics research is greatly spurred on by new, powerful ‘deep sequencing’ technologies that provide not only complete genome sequences in a matter of days, but also the frequencies of DNA variants and different RNA molecules present in a sample (see Box in [8]).

Inheritance of acquired stress via sperm

Isabelle Mansuy at University of Zurich Switzerland and colleagues studied the effects of traumatic stress in early life on sperm sncRNA in the mouse model of unpredictable maternal separation combined with unpredictable maternal stress (MSUS), which is known to cause behavioural effects across generations [9]. Males showed reduced avoidance and fear in several tests, and these characteristics were transmitted to the offspring. Metabolic changes were also evident. Insulin levels were normal in F0 mice subjected to the stress treatment (referred to erroneously as F1 by authors), but were lower in the subsequent generation F1 mice (referred to as F2 by authors). Blood glucose was normal in F0 animals but also lower in F1 mice, both at baseline and following an acute restraint stress. Furthermore, F0 MSUS males had normal baseline glucose levels and clearance on a glucose tolerance test, but showed a larger decline in blood glucose on an insulin tolerance test. F1 MSUS animals had normal glucose at baseline but lower glucose rise on glucose tolerance test, and normal glucose decrease on insulin tolerance test. These anomalies suggest insulin hypersensitivity. F1 but not F0 also showed hypermetabolism, as their body weight was lower than that of controls despite high caloric intake. The alterations were overall more marked in F1 mice, probably because the (lingering) effects of stress are present starting at conception, whereas they only occurred after birth in F0 mice.

Deep sequencing revealed that several miRNAs were up-regulated in F0 MSUS mice, and 73 potential miRNA target genes involved in DNA and RNA regulation, epigenetic regulation or RNA binding and processing were identified [5]. PiRNAs were also affected; particularly cluster 1110, which was down-regulated in MSUS sperm. Reverse transcription quantitative PCR confirmed that miR-375-3p, miR-375-5p, miR-200b-3p, miR-672-5p and miR-466-5p were up-regulated in F1 MSUS sperm. MiRNAs were also altered in serum, and in the hippocampus and hypothalamus, brain structures involved in stress response in adult F0 MSUS animals. Moreover, miRNAs were affected in the serum and hippocampus of adult F1 MSUS mice, but not in F1 sperm. MiRNAs were normal in F2 animals (referred to by authors as F3). But F2 MSUS mice have behavioural symptoms similar to those of F0 and F1 animals. It is possible, therefore, that the changes in miRNAs that initially occurred in sperm cells as a result of MSUS are transferred to other non-genomic or epigenetic marks, such as DNA methylation, or histone post-translational modification for maintenance and further transmission (see above).

Nevertheless, miR-375 has been implicated in stress response and metabolic regulation; and mimicking the effect of stress by injecting corticosterone in vivo increased miR-375 expression. One of miR-375 targets is catenin b1 (Ctnnb1), a protein implicated in stress pathways. Cultured cells transfected with miR-375 mimic showed down-regulation of Ctnnb1. Consistent with this, Ctnnb1 was decreased in F2 MSUS hippocampus.

Microinjecting RNAs purified from sperm of MSUS males into wild-type fertilized mouse oocytes caused comparable behavioural and metabolic/molecular effects in the adult males developed from the injected oocytes. Not only that, the offspring of MSUS-RNA-injected mice also showed depressive like behaviours. However, the precise molecular mechanisms involved in sperm-mediated inheritance of stress response remains to be elucidated.

Oliver Rando at University of Massachusetts Medical School Worcester in the United States proposes that rather than thinking sperm carries information about tens or hundreds of thousands of important environmental variables, different environmental conditions, such as toxic substances, social status, etc., simply alter sperm quality, which then affects phenotypes [6]. Thus, most studies in mammals have focussed on different phenotypes such as metabolism in response to paternal diet, behaviour in response to paternal social defeat, and so on. But it turns out that overlapping phenotypes are found in response to distinct paternal treatments.  For example, not only do endocrine disruptors affect future reproductive success of males, so does ancestral exposure to high fat diet in utero. Altering early embryonic development can have effects similar to those observed in paternal environmental exposure. Humans born after in vitro fertilization exhibit altered glucose tolerance. 

Sperms may simply be transmitting some overall stress measure. Several cases of trans-generational effects turn out to affect epigenetically sensitive reporter genes that were not present in the parent subject to genetic or environmental stressors.

Nevertheless, sperm does mediate the transfer of specific genes.

Sperm-mediated gene transfer known since 1971

Sperm-mediated gene transfer has been known at least since 1971. Simian virus SV 40 absorbed on rabbit spermatozoa but failed to penetrate inside. In contrast, sperms exposed to SV40 DNA ended up with SV40 DNA in the posterior part of the acrosome inside the sperm [10]. When these sperms carrying SV40 DNA were fused with CV-1 African green monkey kidney cells, infectious SV40 virus was isolated. When inseminated into rabbits, both unfertilized and one and two cell fertilized ova were obtained, which, when co-cultivated with CV-1 cells gave rise to infectious virus. This was the first evidence that a virus genome can be incorporated into a mammalian spermatozoon and subsequently carried into an ovum during fertilization.

In 1989, mature mouse sperm cells incubated in an isotonic buffer with cloned DNA, was reported to capture DNA molecules over a 15 min period [11]. This was passed on to the F1 progeny by both males and female parents in both Mendelian (classical chromosomal inheritance) and non-Mendelian (non-classical genetics) manner.

Very similar results were obtained with Xenopus laevis sperm cells, whereas egg cells incubated with DNA were unable to take up the DNA.

Only living sperm cells could take up DNA and the temperature makes little difference between o and 37 °C. The DNA rapidly associated with the sperm cells within 10-15 minutes and 50 % of the DNA were stable to extensive washing. No morphological alterations were caused by the DNA uptake. Most of the radioactively labelled DNA ended up in a specific portion of the sperm head.

Extrachromsomal inheritance mediated via sperm

However, foreign sequences transferred by sperm are not often stably integrated, but exist as extrachromosomal structures. A prominent role is played by an endogenous reverse transcriptase of retrotransposon origin. Mature spermatozoa are naturally protected against foreign nucleic acid molecules; but this protection can be abolished under particular environmental conditions, such as those occurring during assisted reproduction.

Over 70 reports of successful sperm-mediated gene transfer (SMGT) had appeared by 2005 (reviewed in [12]). Most provide evidence of post-fertilization transfer and maintenance of transgenes, and several also report subsequent generation of viable F0 animals, the cells of which contain exogenous DNA sequences. In some cases evidence is provided of transmission to the progeny F1 and beyond.  The species include rabbit, sea urchin, mouse, farm bull, insects chicken, pig, human, boar, zebrafish, salmon, loach, various mammals, frog, abalone, rat, carp, rhesus monkey, honey bee, sea bream, and Rohu fish.

The binding of exogenous DNA molecules to sperm cells and subsequent uptake is thought to involve specific processes and factors. When the DNA reaches the nuclear scaffold, a small fraction eventually undergoes recombination with the sperm chromosomal genome at a few selected sites, such as a set of chromatin sites with a distinctive conformation, accessible to nuclease digestion and complexed with histones rather than protamines.

Interaction with exogenous molecules triggers an endogenous reverse transcriptase (RT) in sperm cells. Such RT activity is able to reverse transcribe exogenous RNA molecules – specifically the human poliovirus RNA genome – into cDNA copies, which are transferred to embryos following in vitro fertilization. Retrotransposon/retroviral machinery is thought to be involved in SMGT.

RT was historically associated with the replication of retroviruses and later found to be also encoded by two major classes of repeated elements in higher eukaryote genomes, retrotransposons such as those of the LINE (long intersperse nuclear element) family, and endogenous retroviruses, collectively termed retroelements. As much as 45 % of human and 37 % of mouse genomes consist of retroelements traditionally regarded as useless junk DNA. However, growing evidence is suggesting that endogenous RT plays a role in reshaping and rearranging genomes during normal development and in response to environmental challenges (see [7, 8, 13, 14] (Evolution by Natural Genetic Engineering, Non-Coding RNA and Evolution of Complexity, SiS 63). RT is responsible for numerous genome alterations and acts as a major driving force in evolution. It is expressed in elevated levels in embryos, embryonic tissues and tumours. In contrast, RT expression is low to zero in terminally differentiated cells. The mammalian genital tract, germs cells and gametes are other preferential sites of retroviral/retrotransposon genes. Endogenous RT is preferentially expressed in tissues with a high proliferation potential and may therefore be involved in regulating cell growth and differentiation. Consistent with this hypothesis, inhibition of endogenous RT arrests embryo development in early pre-implantation stages, and also modulates proliferation and differentiation in transformed cell lines.

The RT activity detected in mouse sperm cells is most likely encoded by sperm chromatin organized in an active nucleohistone conformation and enriched in LINE-1 sequences. The existence of an active RT in spermatozoa is unexpected. To determine whether foreign RNA can be a suitable substrate for the sperm RT, an IVF experiment was carried out in which sperm cells were incubated with a RNA vector marked with a b-galactosidase gene. The b-galactosidase expressing RNA vector was taken up by mouse sperm cells, reverse-transcribed and delivered to embryos upon IVF and propagated in founders in mosaic fashion, and were passed on likewise to the F1 progeny [15]. This phenomenon is referred to as “sperm-mediated reverse gene transfer.” The population of cDNA molecules that are reverse transcribed exhibits peculiar features that distinguish them from other transgenes. First these sequences are maintained as low copy number (<1 per genome). Second they show a mosaic distribution in founder animals, and are sexually transmitted from founders to F1 progeny, where they are again propagated in mosaic fashion as extrachromosomal episomes. These features suggest they are not integrated in the host genome but replicated independently and nevertheless able to enter germ cells and/or become replicated within germ cells during gamete formation and before maturation.

In vitro studies show that exogenous nucleic acid uptake by mature sperm is strongly inhibited by IF-1, an abundant glycoprotein in the seminal fluid of mammals and on the sperm surface in marine species. In mammals, IF-1 is probably lost during the movement of sperm into the female genital tract. In aquatic species, sperm cells are exposed to exogenous nucleic acids derived from cell catabolism of animal and plant origin present in the seawater. However, the exposure of sea urchin sperm to a low ionic strength medium detaches IF-1, thereby allowing exogenous nucleic acid molecules to interact with the sperm.

Similar conditions may arise naturally as for example in sea water near the mouth of a river.

After the nucleic acids are taken up they may be degraded by endogenous nucleases. Surviving nucleic acids may either integrate into the sperm nucleus or more likely, remain as episomes.

In humans, IVF may inadvertently lead to introduction of foreign nucleic acids into sperm cells and delivered to egg cells in fertilization. They may even transfer HIV and other viral particles.

Genome integration depends on route of delivery

Corrado Spadafora at Istituto Superiore di Sanita Rome Italy , who rediscovered the SMGT in the late 1980s and publicised it widely, wrote in 2008 [16]:“It is now widely accepted that spermatozoa of virtually all animal species have the spontaneous ability to take up exogenous DNA molecules and to deliver them to oocytes at fertilization.” However the subsequent fate of sperm-bound DNA after delivery to the oocyte is still contradictory, and appears to depend on the procedures use.

The inheritance of non-integrated episomal structures is highly probable when the foreign DNA molecules are directly incubated with intact sperm then used in fertilization. Integration in the host genome appears to be favoured with protocols that avoid a direct interaction between exogenous nucleic acid molecules and the sperm cell membrane, as by incubating sperm cells without intact cell membrane with DNA molecules followed by microinjection of the sperm-DNA complex into oocytes, or DNA slipping through the membrane wrapped in lipids, or incubating sperm cells with antibody directed against specific proteins expressed on the sperm surface.

The internalized exogenous DNA sequences reach the nuclear scaffold of sperm cells, where they are subjected to rearrangement(s) by endogenous nucleases and undergo recombination events that eventually cause their integration into the sperm genome. A genome library from murine sperm cells incubated with a reporter plasmid and screened to identify the integration sites suggests that integration occurs in one or only very few preferential site(s) in the sperm genome, and is a very infrequent event.

The nucleohistone fraction of mouse sperm chromatin has features closely resembling those of active chromatin in somatic cells in being nuclease sensitivity, in its nucleosome organization, and low levels of methylation in the DNA. Sequence analysis of randomly isolated clones from that fraction suggested an unexpected enrichment in DNA sequences of retrotransposons; and among the most abundant are reverse transcriptase encoding LIN-1 ORF2 sequences. To test whether a functional RT activity is present in mature sperm, human poliovirus chromosomal RNA was used as it replicates through a RNA (-) strand with no DNA intermediate (as is the case for retroviruses). The poliovirus RNA was indeed taken up by the sperm cells, the nucleic acid-binding molecules present on the sperm cell surface did not show any obvious preference in their interaction with DNA or RNA. The poliovirus was reverse transcribed into cDNA and transferred to oocytes during IVF and further transmitted to two-cell embryos. Moreover, electron microscope analysis using anti-RT antibody showed that RT molecules were stably associated with the sperm nuclear scaffold.

To test whether newly synthesized cDNAs in sperm behave as biologically active retrogenes, sperm cells were incubated with RNA populations transcribed from a construct expressing a b-galactosidase reporter gene and then used in IVF. A founder F0 progeny, and a subsequent F1 progeny by normal breeding were produced. Direct PCR analysis of DNA samples from both F0 and F1 animals confirmed that b-gal containing cDNAs were generated in sperm, delivered to oocytes, and propagated in mosaic fashion throughout embryogenesis in various tissues of adult animals, and transmitted likewise to the next generation. The enzyme was expressed in a variety of positive tissues in both Fo and F1 animals.

RT-dependent process is triggered not only when sperm cells are incubated with RNA but also when they are incubated with DNA. Sperm cells were incubated with a retrotransposition cassette containing a DNA construct with an enhanced green fluorescence protein (EGFP) reporter gene interrupted by a g-globin intron placed in the opposite orientation to that of EGFP transcription. In order to be expressed, the reporter gene needed to go through a reverse transcription step. The exogenous DNA was taken up by the sperm and into the nuclei where it was transcribed; the primary RNA transcript was spliced and finally reverse-transcribed into EGFP-containing cDNA copies. Interestingly, only a small proportion of the newly synthesized cDNAs remained in the sperm head. Most of the reverse transcribed molecules were released into the incubation medium where they were available for further interaction with sperm cells. This led to a steady-state in which the vast majority of the sperm cell population was associated with foreign cDNAs. These reverse transcribed cDNA delivered to oocytes at fertilization exhibited the same features as those obtained when sperm cells were incubated with exogenous RNA, essentially as extrachromosomal, low-copy structures. The EGFP reporter gene was found to be expressed in the vascular epithelium of various positive tissues of adult animals.

In a new study, mice xenografted with human melanoma cells stably expressing a EGFP-encoding plasmid [17] released EGFP RNA from the xenografted human cells into the bloodstream and eventually into the spermatozoa of the mice via exosomes. Exosomes are RNA containing vesicles released by cells into the bloodstream as part of a nucleic acid intercommunication system (reviewed in [18]).

Soma to germline genome transmission

However, soma to germline genome transmission may not be such an infrequent event in the course of evolution. This proposal was first put forward for the immune system by immunologist Ted Steele in 1979 [19], and refined in subsequent publications (reviewed in [20]). Essentially, it invokes the clonal expansion of antibody-selected B lymphocytes expressing hypermutated high affinity antibodies providing the source of abundant RNA transcripts that are reverse transcribed into cDNA and integrated into the germline genome.

One strong candidate for such soma to germline genome transmission is the immunoglobulin heavy chain variable (IGHV) region which consists of about 100 V genes in the germline. This region is so diverse that it is believed no two chromosomes share the same set of IGHV gene segments in the entire human population [21]. Steele points out that as hypermutation of the V region only occurs somatically in B cells, such diversity can only result from a soma to germline genome transmission in the course of evolution. There are indeed mutational changes indicative of RNA editing and DNA deamination associated with hypermutations in somatic B cells.

Implications for evolution and safety of artificial genetic modification

As in the numerous mechanisms of natural genetic modification reviewed in other articles of this series [7, 8, 13, 14, 22] (Horizontal Transfer of GM DNA Widespread, SiS 63), sperm-mediated inheritance of acquired characters greatly speeds up evolution, and at the same time, greatly multiplies the hazards of artificial genetic modification. Epigenetic toxicology is a new discipline [3] telling us in no uncertain terms that exposure of one generation to new toxins and nucleic acids introduced into our food chain from GM food and feed can be passed on and amplified in subsequent generations. We are risking not just our own lives but the well-being of our children, grandchildren, and great grandchildren.

Article first published 14/07/14


  1. Ho MW. Epigenetic inheritance, “what genes remember”. Science in Society 41, 4-5, 2009.
  2. Ho MW. Caring mothers trike fatal blow against genetic determinism. Science in Society 41, 6-9, 2009.
  3. Ho MW. Epigenetic toxicology. Science in Society 41, 13-15, 2009.
  4. Anyway MD, Cupp AS, Uzumcu M and Skinner MK. Epigenetic transgenerational actions of endocrine disruptors and male fertility. Science 2005, 308, 1466-8
  5. “The epigenetics heretic”, Jocelyn Kaiser, News Focus, Science 2014, 343, 3610-3.
  6. Rando OJ. Daddy issues: paternal effects on phenotype. Cell 2012, 151, 702-8.
  7. Ho MW. RNA inheritance of acquired characters. Science in Society 63 (to appear) 2014.
  8. Ho MW. Nucleic acid invaders from food confirmed. Science in Society 63 (to appear) 2014.
  9. Gapp K, Jawaid A, Sarkes P, Bohacek J, Pelczar P, Prados J, Farinelli L, Miska E & Mansuy IM. Implications of sperm RNAs in transgenerational inheritance of the effects of early trauma in mice. Nature Neuroscience 2014, published online 13 April doiL10.1038/nn.3695
  10. Brackett BG, Baranska W, Sawicki W, Koprowski H. Uptake of heterologous genome by mammalian spermatozoa and its transfer to ova through fertilization. Proc Natl Acad Sci USA  1971, 68, 353-7.
  11. Lavitrano M, Camaioni A, Fazio VM, Dolci S, Farace MG and Spadafora C. Sperm cells as vectors for introducing foreign DNA into eggs: genetic transformation of mice. Cell 1989, 57, 717-23.
  12. Smith K and Spadafora C. Sperm-mediated gene transfer: applications and implications. BioEssay 2005, 27, 551-6.
  13. Ho MW. Evolution by natural genetic engineering. Science in Society 63 (to appear) 2014.
  14. Ho MW. Non-coding RNA and evolution of complexity. Science in Society 63 (to appear) 2014.
  15. Sciamanna I, Barberi L, Marlire A, Pittoggi C, Beraldi R et al. Sperm endogenous reverse transcriptase as mediator of new genetic information. Biochem. Biophys Res Commun 2003, 312, 1039-46.
  16. Spadafora C. Sperm-mediated ‘reverse’ gene transfer: a role of reverse transcriptase in the generation of new genetic information. Human Reproduction 2008, 735-40.
  17. Cossetti C, Lugini L, Astrologo L, Saggio I, Fals S and Sadafora C. Soma-to-gerline transmission of RNA in mice xenografted with human tumour cells: possible transport by exosomes. Plos One 2014 (in press).
  18. Ho MW. The new genetics and natural versus artificial genetic modification. Entropy 2013, 15, 4748-81.
  19. Steele EJ. Somatic Selection and Adaptive Evolution; On the Inheritance of Acquired Characters, Williams-Wallace, Toronto, 2nd edition, University of Chicago Press, Chicago, 1981.
  20. Steele EJ. Lamarck and Immunity: somatic and germline evolution of antibody genes. J Roy Soc W Aust 2009, 92, 437-46.
  21. Pramanik S, Cui X, Wang H-Y, Chimge N-O, Hu G, Shen L, Gao R and Li H. Segmental duplication as one of the driving forces underlying the diversity of the human immunoglobulin heavy chain variable gene region. BMC Genomics 2011, 12, 78 (12 pp).
  22. Ho MW. Horizontal transfer of GM DNA widespread, but no one is looking, almost. Science in Society 63 (to appear) 2014.

Got something to say about this page? Comment

Comment on this article

Comments may be published. All comments are moderated. Name and email details are required.

Email address:
Your comments:
Anti spam question:
How many legs on a tripod?