Zscan4: The Possible Basis of ALT

To develop an unbreachable defense against cancer, SENS Foundation is pursuing the WILT (Wholebody Interdiction of Lengthening of Telomeres) strategy (OncoSENS) of systematically deleting genes essential to the cellular telomere-maintenance mechanisms (TMM) from all somatic cells, while ensuring ongoing tissue repair and maintenance through periodic re-seeding of somatic stem-cell pools with autologous TMM-deficient cells whose telomeres have been lengthened ex vivo. In addition to the deletion of one or more genes coding for essential element(s) of the telomerase holoenzyme, success will also require the deletion of some essential element of the machinery for the Alternative Lengthening of Telomeres (ALT) phenomenon, obvserved in a minority of cancer cells.

Heretofore, the identity of that machinery has been elusive. Yeast cells have the ability to lengthen telomeres through a telomerase-independent mechanism involving telomere recombination, and there has been evidence for some time suggesting that ALT cancers lengthen telomeres through a similar process. Intriguingly,  some cells derived from mice lacking the telomerase RNA component (TERC) (embryonic stem cells (ESC) and subpopulations of somatic cells (fibroblasts, splenocytes proliferating in response to immunization)) exhibit limited telomere maintenance through a mechanism also apparently involving recombination (as telomere sister-chromatid exchange (T-SCE)).(1-3) Additionally, murine ESC genetically deficient for some DNA methyltransferases exhibit  elongated telomeres, accompanied by increased telomeric recombination  and an ALT-like phenotype.(4)

Based in part on findings in B-cells from TERC-/- mice (2), SENS Foundation funded preliminary explorations by Mathias Bollmann (Universitätsklinikum Hamburg-Eppendorf, Hamburg) to test the possibility that this recombination might be accomplished through activation-induced cytidine deaminase (AID), a gene essential to immunoglobulin somatic hypermutation and class switch recombination and sharing some known characteristics of ALT. However, the results did not support AID as a likely ALT candidate (personal communication, M. Bollmann), and the search continued.

In 2007, Liu et al reported that

oocytes actually have shorter telomeres than somatic cells, but their telomeres lengthen remarkably during early cleavage development. Moreover, parthenogenetically activated oocytes also lengthen their telomeres, thus the capacity to elongate telomeres must reside within oocytes themselves. Notably, telomeres also elongate in the early cleavage embryos of telomerase-null mice, demonstrating that telomerase is unlikely to be responsible … Coincident with telomere lengthening, extensive … (T-SCE), and colocalization of the DNA recombination proteins Rad50 and TRF1 were observed in early cleavage embryos. Both T-SCE and DNA recombination proteins decrease in blastocyst stage embryos, whereas telomerase activity increases and telomeres elongate only slowly.(5)

Now, a new report currently in advance publication in Nature appears to identify the genetic basis of this phenomenon.

A Zscan4 (zinc finger and SCAN domain containing 4) gene cluster includes six transcribed paralogous genes (Zscan4a–f) that share high sequence similarities and are thus collectively called Zscan4. A sharp expression peak of Zscan4 marks the late two-cell stage of mouse embryos and is essential for embryo implantation and blastocyst outgrowth in tissue culture. Zscan4d is transcribed predominantly in two-cell embryos, whereas Zscan4c is transcribed predominantly in ES cells and is associated with self-renewal. … Here we show Zscan4 to be essential for long-term culture of ES cells and maintenance of karyotype integrity associated with regulated telomere recombination in normal undifferentiated ES cells. … Only 5% of ES cells express Zscan4 at a given time, but nearly all ES cells activate Zscan4 at least once during nine passages.(6)

Based on Zscan4 expression tracked by GFP and fate-mapping studies of ESC manipulated by transient activation of Zscan4, it appears that this 5% represents an equilibrated state, the net effect of ~3% of ESC initiating expression of the gene on a given day, even as ~50% of previously Zscan4-expressing cells cease expression.

The transient Zscan4-positive state is associated with rapid telomere extension by telomere recombination and upregulation of meiosis-specific homologous recombination genes, which encode proteins that are colocalized with ZSCAN4 on telomeres. Furthermore, Zscan4 knockdown shortens telomeres, increases karyotype abnormalities and spontaneous sister chromatid exchange, and slows down cell proliferation until reaching crisis by passage eight. …  Surprisingly, Zscan4 induction led to elongated telomeres in Zscan4- rescue cells. The telomere length distribution … indicated that the average increase in telomere length was due to an overall shift and not due to a few abnormally long telomeres. …. The telomere lengthening by Zscan4 was not associated with increased telomerase activity, as shown by telomeric repeat amplification protocol (TRAP) assay, nor abolished by eliminating telomerase activity with TERT [telomerase reverse transcriptase] knockdown, suggesting a telomerase-independent mechanism for the Zscan4 activity. … 

As Zscan4 is a common marker for ES cells and the two-cell embryos, we investigated whether Zscan4 activated the T-SCE apparatus … As expected, the frequency of T-SCE in all control cells was low … By contrast, Zscan4c induction … for 3 days resulted in >10-fold more T-SCE events in 76% of the nuclei tested. … T-SCE mediated by Zscan4 is not associated with an increase of general SCE, and the normal karyotype remains stable with a lower spontaneous SCE rate … Therefore, we conclude that transient expression of Zscan4c promoted telomere recombination, leading to telomere elongation.(6)

Collectively, these data provide a strong basis for investigating the possibility that Zscan4 (or one or more of its paralogs) is the genetic basis of the ALT phenotype. It is surprising, in fact, that the authors only peripherally touch on this possibility and its therapeutic implications, rather than making them a central focus of discussion. They do caution that

unlike cells usually associated with T-SCE, such as survivors of telomerase knockout TERC-/- ES cells and tumour cells lacking telomerase reactivation [ie, ALT cells], … telomerase is active in undifferentiated ES cells expressing Zscan4. We have also shown that Zscan4-mediated telomere extension does not require telomerase; however, it remains unclear whether Zscan4-mediated telomere extension can compensate the telomere loss in TERC-/- ES cells, which cease to proliferate after .>450 population doublings.(6)

… but such limitations may very reasonably hypothesized to be the imposed by the tight physiological regulation of the gene cluster, which they have demonstrated to allow only a transient window of intense TMM activity before turning ongoing telomere maintenance over to telomerase. But such regulatory limitations are precisely the kinds of controls that are broken down by selective pressure in the intensive evolutionary engine of cancerous and precancerous cells. Other investigators will doubtless follow up on this lead, interrogating ALT cancer cell lines and telomerase-deficient mice to see if Zscan4 is indeed responsible for telomere maintenance in these systems.

Should Zscan4 prove to be responsible for the ALT phenomenon, there will be further interest in developing small molecules and other strategies to inhibit Zscan4 in the ~10-15% of cancers in which ALT is implicated.  If so, then the additional research priorities for WILT become clear. 

First,we will need to determine whether ALT may play some physiological role in adult somatic cells. To date, the limited reports available suggest that Zscan4 is only physiologically active in ESC, and the remarkable genomic stability of ESC compared to both somatic cells and the germ line are consistent with the confinement of Zscan4 activity to this narrow window of early development. If indeed there is no such role, then its deletion in somatic cells should be harmless, and the path ahead relatively straightforward. In one scenario, cells intended for eventual therapeutic use in renewal of tissue stem cell pools would have the gene cluster deleted from the outset, but transiently introduced into early precursors through DNA plasmids — perhaps along with TERT for the parallel purpose. In a less likely scenario, ESC-equivalent cells would be derived with an intact Zscan4 locus, and then advanced to the tissue-specific stem cell stage before final deletion. Either way, Zscan4 would have fulfilled its role in telomere maintenance and genomic stability during the earliest stages of cell derivation, but would no longer be present in cells actually delivered to patients. At a later stage in the maturation of the intervention, native Zscan4 would be systematically deleted in vivo through somatic gene therapy.

On the other hand, should Zscan4 itself prove to have some residual physiological role in the soma, then it may be that this role is independent of its involvement in telomere maintenance, as now seems likely in the case of some putative physiological roles of  TERT. If so, then investigators may proceed by probing other genes in its network, to see what other elements might achieve the goal of rendering the derived cells ALT-incompetent while preserving Zscan4’s non-TMM function. Another possibility is that only one or a few paralogs of Zscan4 may have this property, and the deletion of other elements would allow the relevant function to continue unperturbed while denying nascent cancer cells a TMM. This is parallel to the case of telomerase itself, where the goal seems likely to be achieved by deleting TERC while leaving TERT intact.

Another way that this finding may prove useful for therapeutic use in rejuvenation engineering comes from the observation that during Zscan4 expression, ESC are positive for the transcription factor  Oct4, now famously necessary to the pluripotency and self-renewal of ESC,  and widely-known today as a result of its central place in the cocktail of factors used for induced cellular reprogramming. Aside from reinforcing the view that cells are yet in an undifferentiated state during Zscan4 expression,  the authors note that “the expression level of Zscan4 in induced pluripotent stem (iPS) cells is comparable to ES cells, suggesting that iPS cells may have regained the ability to undergo ES-like genome maintenance [and, of course, to have done so through this mechanism -MR]. By selecting cells able to activate Zscan4, it may be possible to enrich cultures with cells more suitable for future therapeutic purposes” — a very useful property, as the low efficiency of reprogramming remains a significant hurdle to the eventual therapeutic use of iPS-derived cells and engineered tissues for repair and rejuvenation of aging tissues (RepleniSENS).

Overall, this understated and cautious report is very promising in its implications, and may well come to be comemmorated as a landmark paper in the eventual development of an ultimate cure for cancer, and thus for its contribution to a comprehensive panel of biomedical solutions for age-related disease, disability, and death.