Decreased, atRA concentration leads to Oocyte meiosis, disrupted

Mouse deletion models for the atRA synthesis enzymes Aldh1a1, Aldh1a2 and Aldh1a3 showed decreased expression of Stra8 in double (Aldh1a2/3) and triple (Aldh1a1/2/3) knockouts, although ultimately some germ cells were observed undergoing meiosis in these ovaries, suggesting that atRA is not essential for meiotic onset or progression(Chassot et al, 2020; Kumar et al, 2011). Similarly, transgenic mice lacking the three atRA nuclear receptors (RAR-α, -β, -γ) showed reduced levels of Stra8, although ultimately some germ cells were observed undergoing meiosis and were capable of producing live offspring (Vernet et al, 2020). Whether or not these models led to impaired fertility (such as sub-fertility) has not been elucidated and the size of their oocyte pools were not determined. In addition, the completeness of the genetic deletions in these models is not clear (discussed in (Spiller & Bowles, 2022)).

Gain of function mouse ovary models for CYP26A1 and CYP26B1 show that CYP26B1 can prevent oocytes from entering meiosis (as assessed by failure to induce Stra8 expression), whereas CYP26A1 does not have the same effect despite being a potent atRA degrading enzyme. This suggests that CYP26B1 works by additional mechanism(s) other than RA degradation (Bellutti et al, 2019).

In vitro and ex vivo, it has been conclusively shown that low levels (as low as 1uM) of exogenous atRA can induce germ cells to enter meiosis in mice (Bowles et al, 2010) and rats (Livera et al, 2000) and, similarly, that it is necessary to achieve meiosis in in vitro-derived oocytes via primordial germ cells (PGCs)/PGC-like cells (PGCLCs) (Miyauchi et al, 2017). Yet, its exact role in vivo is under debate.

Whilst the relative levels of endogenous atRA produced by the ovary (for any species) remains unknown, similarly, the quantitative relationship between atRA levels and induction of meiosis also remains unclear. As such, the quantitative understanding of how much atRA needs to be reduced to prevent germ cells to enter meiosis in vivo is rated low.

The time-scale for this KER is relatively short, limited to just a couple of days in e.g. mouse models. The induction of meiosis occurs shortly after the germ cells have colonized the ovary and occurs asynchronously (Bullejos & Koopman, 2004) (in mice this begins at E13.5 and is completed for all germ cells 2 days later at E15.5). Proliferation is halted and cells progress through leptonema, zygonema, pachynema, and arrest in diplonema of prophase I prior to birth (Zamboni, 1986). Time and duration of oogenesis varies between species, with rats the shortest duration of only 1-2 days, with other mammals such as pigs, cows, monkeys and humans lasting months (Peters, 1970).

The rat model of vitamin A deficiency (VAD) revealed severe defects to meiosis induction when Vitamin A was restricted/removed from the diet at E10.5, which is just 3 days prior to normal meiotic induction (Li & Clagett-Dame, 2009). Shorter time-frames have not been assessed to date, nor has rescue of VAD during later embryonic time-points been attempted.

During development, retinoic acid homeostasis is regulated by feedback loops, as both too much and too little RA can have deleterious effects on the embryo or fetus. The availability of atRA is regulated locally by maintaining a balance between synthesis (ALDH1 enzymes) and metabolism (CYP26 enzymes) (Kedishvili, 2013; Niederreither & Dollé, 2008; Roberts, 2020; Teletin et al, 2017).

The expression of Aldh1a2 and Cyp26a1 can act as part of a negative feedback loop in response to changes in RA levels. Exogenous atRA suppresses expression of Aldh1a2 (Niederreither et al, 1997) whereas blocking atRA signalling increases expression of Aldh1a2. Although Cyp26 expression does not require atRA, addition of atRA greatly increases the expression of Cyp26a1, and conversely, reduced levels of atRA reduces Cyp26a1 expression (de Roos et al, 1999; Hollemann et al, 1998; Ross & Zolfaghari, 2011; Sirbu et al, 2005). Negative feedback loops also extend to the enzymes that convert retinol to all-trans retinaldehyde as well as other related enzymes (Feng et al, 2010; Strate et al, 2009), including Ski, which seem to have cell-type specific roles (Melling et al, 2013; Niederreither & Dollé, 2008).