Intergenerational Transmission of Metabolic Changes in Oocytes From Aged Mice

ABSTRACT

The increase in maternal age of pregnancies is a global phenomenon that may have wide‐ranging implications for the future health of the next generations. We have previously shown that oocytes from females at advanced maternal age (AMA F0) accumulate intracytoplasmic lipid droplets (LDs), and that oxidative changes to lipids in oocytes from AMA F0 mice are maintained in preimplantation embryos. Here we explore whether oxidative changes are transmitted to the foetus, and what effects this has on neonatal and adult organ development, and the transgenerational inheritance of these changes. First, we show increased antioxidants in lipid‐rich organs (liver and brain) of AMA‐derived prenatal mice (AMA F1), indirectly showing increased oxidative stress. Then, we provide evidence of metabolic reprogramming in adult offspring of AMA and the accumulation of lipids in AMA‐derived third generation (AMA F3) mouse oocytes. In particular, we demonstrate the accumulation of retinoids and other mediators of oxidative phosphorylation (OXPHOS) in AMA F0 and AMA F3 oocytes. Altogether, an altered oxidative metabolism of AMA F0 oocytes may constitute a base of oxidative changes in the organs of offspring and of their transgenerational inheritance to AMA F3 oocytes. Our findings indicate a remodelling of lipid homeostasis in oocytes of female mice derived from AMA great‐grandmothers and highlight the need to take a closer look at the inheritance of metabolic changes from mothers of advanced age into their offspring.

This study shows that oxidative changes in oocytes from aged female mice lead to lipid droplet accumulation and altered energy and retinoid metabolism. These changes persist in embryonic organs, triggering antioxidant responses in lipid‐rich tissues. Remarkably, altered profiles are transmitted across generations, reaching third generation, suggesting that maternal aging remodels oocyte metabolism with transgenerational effects.

Abbreviations

AEA

anandamide

AMA

advanced maternal age

ART

assisted reproductive technologies

BMI

body mass index

CARS

Coherent anti‐Stokes Raman spectroscopy

Cer

ceramides

Chol

cholesterol

DOHaD

Developmental Origin of Health and Disease

E16

embryonic day 16

E19

embryonic day 19

ER

endoplasmic reticulum

ET

embryo transfer

F:P

Foetal‐to‐placental weight ratio

FTIR

fourier transform infrared spectroscopy

GSH

glutathione

GSSG

glutathione oxidised

GV

germinal vesicle

HCA

hierarchical clustering analysis

Hchol

hydroxylated cholesterol

LDs

lipid droplets

mtDNA

mitochondrial DNA

NAD⁺

nicotinamide adenine dinucleotide

NAFLD

non‐alcoholic fatty liver disease

NP

natural pregnancy

OS

Oxidative stress

OXPHOS

oxidative phosphorylation

PC

phosphatidylcholine

PCA

principal component analysis

PC O‐

ether‐linked phosphatidylcholine

PE

phosphatidylethanolamine

PI

phosphatidylinositol

PND

post natal day

PPAR

peroxisome proliferator–activated receptors

PS

phosphatidylserine

PUFA

polyunsaturated Fatty Acids

ROS

reactive oxygen species

SAM

S‐adenosylmethionine

SFA

saturated fatty acids

SM

Sphingomyelins

Sph

sphingolipids

SREBP‐1

sterol regulatory element‐binding protein 1

Sulf

sulfatides

UFA

unsaturated fatty acids

YMA

young maternal age

α‐KG

alpha‐ketoglutarate

1. Introduction

It is well known that advanced maternal age (AMA) is a major risk factor for pregnancy success and offspring health, but less is known about the transgenerational effects of AMA pregnancies. This is a growing area of research, with multiple avenues remaining as yet unexplored. In particular, the mechanisms behind metabolic alterations in offspring and grandoffspring from AMA pregnancies have not been adequately explored, despite abundant evidence confirming the heritability of metabolic programming (Dabelea and Crume 2011; Harville et al. 2017).

As more evidence comes to light, age‐related lipid oxidation damage has been found to be increased in both human and mouse ovarian follicular cells (Lim and Luderer 2011; Smits et al. 2023). This has led to concerns that antioxidant systems can be overwhelmed in AMA offspring, starting even before conception: the offspring of aged mothers often have reduced antioxidant defenses, making them more susceptible to oxidative stress (Tarín 1996).

AMA pregnancies have therefore been implicated in the generation of dysfunctional mitochondria, and their transmission to offspring (Smits et al. 2023). Perhaps most importantly with regards to AMA, mitochondria are passed on via the female line, presenting one pathway for AMA alterations to be inherited through multiple generations of offspring. It is known that age‐related mtDNA mutations in oocytes can be inherited by offspring, which in a vicious cycle leads to further mitochondrial dysfunction and lipid accumulation (Zhang et al. 2023; Cai and Chen 2024).

The effects of periconceptional changes on AMA‐derived offspring are unclear, but concerning. In mice, offspring from AMA females are characterised by increased bodyweight, increased blood pressure, decreased spleen weight, and impaired liver performance during the glucose tolerance test (Velazquez et al. 2016). Less is known about the effect of AMA on gamete metabolism. Since advanced maternal age is frequently accompanied by systemic metabolic alterations, which may independently shape the developmental environment, disentangling the contributions of chronological ageing and maternal metabolic status therefore remains a central challenge when interpreting offspring phenotypes.

We previously reported age‐dependent accumulation of intracytoplasmic lipid droplets (LDs) in human and mouse oocytes, most likely as a consequence of oxidative stress (OS) resulting from female ageing (Musson et al. 2022; Bisogno et al. 2024). Moreover, this LD accumulation in AMA oocytes was subsequently maintained during preimplantation development into blastocysts. However, it is important to distinguish indirect indicators of oxidative imbalance from direct molecular markers of oxidative damage, which remain comparatively underexplored in AMA‐derived embryos.

It is not known whether AMA‐derived blastocysts maintain a high lipid level due to epigenetic reprogramming, persistence of OS in the embryo, or because it takes longer for the autonomously developing preimplantation embryo to return to healthy lipid levels. Maternal transfer of resources to the embryo starts with implantation and the initiation of normal placental function. This nutrition provides the embryo with a reliable stream of material to perform antioxidant processes and recover lipid levels to a balanced state. If this supply is compromised or insufficient, lipid‐rich areas of the developing foetus—such as the brain—may be most at risk of damage from lipid peroxidation and oxidative stress (Pham et al. 2023). We thus wanted to know if the signs of OS in AMA‐derived mice are still present at fetal stages and in adulthood. Moreover, we hypothesised that the altered profile of lipids may be transmitted transgenerationally through the germ line.

As ageing‐related lipid accumulation in oocytes is a phenomenon observed across diverse mammalian species, including human (Bisogno et al. 2024), similar transmission of metabolic changes in AMA pregnancies through the female germline is highly probable. Given the importance of lipids as an energy source, and of LDs as a store of damaged lipids, we find it concerning how few studies are published on this topic. Especially given that the oocyte relies on maternally‐supplied lipids and cannot acquire new resources until embryonic implantation, the role of lipids in aged oocytes remains insufficiently understood.

2. Results

2.1. Lipid Droplet Accumulation in Aged Oocytes

Our previous studies have shown an increase in overall lipid content, and LDs specifically, in oocytes of aged females. This seems to be the case in both mice (Musson et al. 2022) and humans (Bisogno et al. 2024), suggesting either a shared mammalian consequence of ageing with negative effects on the germ cells, or an evolutionarily conserved adaptation aimed to improve the success rates of pregnancies at advanced age. We also found that the level of lipids in the oocyte corresponds to the level of lipids in the embryo: delipidated oocytes and aged oocytes will generate blastocysts with low or high levels of lipids, respectively (Bisogno et al. 2024).

To investigate this matter in further detail, we performed non‐destructive Raman spectroscopy imaging and Coherent anti‐Stokes Raman spectroscopy (CARS) on oocytes from young and aged (YMA; AMA F0) C57BL/6 females (Figure 1). With these techniques, we are once again able to confirm the accumulation of LDs in aged oocytes, evidenced by an increase in the area of the cell covered by LDs (Figure 1B), and an increase in larger‐sized LDs (Figure 1C). Taken together, these results link the ageing process with lipid accumulation in greater numbers of LDs, and for these droplets to swell to unusual sizes. At present, the role of LDs in germ cells is poorly understood, but there is growing evidence demonstrating a functionality in sequestering damaged lipids and toxic byproducts of ROS damage (Ivanovska et al. 2023; Ovejero et al. 2023). Moreover, we were able to visualise the spatial aggregation of LDs in close proximity to the plasma membrane and nucleus in NP AMA F0 oocytes. This localisation may present an opportunity for peroxidised lipids and toxic ROS products to damage other cellular membrane components, with potential downstream implications for embryo and foetal development (Figure 1A).

FIGURE 1.

Accumulation of LDs in NP AMA F0 oocytes. (A) Raman spectroscopic analysis of oocytes obtained from 3‐month‐old C57BL/6 females (NP YMA) and 12‐month‐old C57BL/6 females (NP AMA F0). Representative Raman images of oocytes obtained by integration in the region 2830–2900 cm⁻¹ (lipids) with representative cluster analysis images of main classes: LDs (red), larger LDs (violet), cytoplasm (orange), lipids+protein (green), nucleus (light blue), and nucleolus (dark blue); (B) Total area by percentage of LDs in NP YMA and NP AMA F0 oocytes (KMC analysis); (C) Increased number of larger LDs in NP AMA F0 oocytes. Left: Representative coherent anti‐Stokes Raman spectroscopy (CARS) images of NP AMA F0 and NP YMA oocytes; Right: Colours associated with each identified LDs size.

Given the close association between ageing and cellular ROS damage, we believe the results presented here support our hypothesis of oxidative stress causing lipid damage and LD accumulation, and provide evidence for the damaging effects of AMA on embryo development. What effects these might have on foetal and pup development was the next focus of our investigations.

2.2. Reduced Growth and Increased Level of Antioxidants in the Liver and Brain of AMA‐Derived Prenatal Mice

Since aged females are at higher risk of OS due to age‐related decline in cellular repair mechanisms and organelle function (Harman 1956), we were interested in investigating whether this redox imbalance passes to foetuses and impacts their development. To address this, we examined lipid‐rich organs of first‐generation foetuses from aged mothers (NP AMA F1), as these organs are highly susceptible to OS and serve as critical indicators of redox imbalance and metabolic dysregulation (Sohal and Orr 2012). We also generated first‐generation foetuses from aged mothers after embryo transfer to young recipient females (ET AMA F1) to determine whether the effects seen here are a result of oocyte‐driven alterations, or a result of gestation in aged uteri.

We observed elevated levels of both GSH (reduced glutathione) and GSSG (oxidised glutathione) in foetal livers from NP AMA F1 at embryonic Day 16 (E16) compared to young controls (young maternal age, NP YMA) (Figure 2A). Similar increases were also detected in ET AMA F1 relative to ET YMA F1, indicating that the altered glutathione status persists despite gestation in a young uterine environment. It suggests that the observed redox imbalance is primarily driven by oocyte‐derived effects of advanced maternal age rather than by gestational environment. This finding implies a compensatory upregulation in antioxidant capacity response, consistent with studies showing that OS triggers upregulation of antioxidant defences as an adaptive mechanism (Sies et al. 2017).

FIGURE 2.

Advanced maternal age alters embryonic redox balance and impairs foetal growth but embryo survival remains comparable. (A) Increased antioxidant levels in foetal liver at Day 16 liver (E16 FL); and (B) in foetal brain at embryonic Day 19 (E19 FB) in NP AMA F1 (upper row) and ET AMA F1 (lower row) in comparison to their respective controls. Mean ± SEM, two‐tailed t‐test; (C) Decreased foetal weight and placental efficiency in NP AMA F1 mice at E19. Mean ± SEM, Two‐tailed t‐test; (D) Decreased brain weight of ET AMA foetuses at E16. Mean ± SEM, Two‐tailed t‐test; (E). Similar embryo survival rate of AMA and YMA embryos after embryo transfer. Chi‐square test *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

We also found a significant decrease in brain weight at E16 ET AMA F1 foetuses (Figure 2D), indicating growth impairment as a result of suboptimal oocyte quality. This is consistent with previous studies demonstrating that OS during foetal development can disrupt cellular proliferation and differentiation, leading to both structural and functional defects in the brain (Slotkin et al. 2005; Dennery 2007; Imanparast et al. 2024). As pregnancy approaches term, OS naturally increases due to heightened metabolic demands (Grzeszczak et al. 2023). However, in NP AMA F1 foetal brains at E19, we observed an abnormal oxidative burden evidenced by higher GSH and GSSG levels compared to young controls (Figure 2B). A similar increase was also observed in ET AMA F1 as well. This suggests that the antioxidant defences in AMA‐derived offspring may be overwhelmed during late gestation, further exacerbating oxidative damage.

We also observed that NP AMA F1 foetuses at E19 showed lower foetal weight, increased placental weight, and a reduced foetal‐to‐placental (F:P) weight ratio compared to young controls, which is indicative of compromised foetal growth and placental efficiency (Figure 2C). These findings are consistent with literature showing that AMA is associated with placental dysfunction via impaired nutrient and oxygen exchange, which can restrict foetal growth (Napso et al. 2019; Guo et al. 2025). The increased placental weight in NP AMA F1 likely reflects a compensatory mechanism to enhance nutrient delivery; however, the reduced F:P weight ratio suggests that the foetus is still not fully nourished. Previous studies have shown that placentae of aged dams are larger in comparison to young dams, with increased markers of OS, thus supporting our finding (Napso et al. 2019). In contrast, ET AMA F1 displayed foetal and placental weights comparable to their young controls, indicating that placental efficiency was somewhat normalised when gestation occurred in a young uterine environment. Notably, embryo survival rates (viable embryos relative to embryos transferred) were comparable between ET AMA and ET YMA groups, indicating that oocyte age does not significantly impair implantation or early embryonic viability. Together, these findings suggest that while advanced maternal age–associated oocyte defects do not compromise embryo survival, they program downstream alterations in foetal growth and placental function.

An interesting observation is that ageing is often accompanied by obesity and other metabolic disorders. We observed significantly higher body weight of aged female mice compared to young counterparts from two different strains and their hybrids (Figure S1). Our mice were fed standard chow, but still gained weight while approaching AMA. This natural progression of increasing body weight, along with the strong association between ageing and metabolic dysfunction, makes it difficult to attribute the oxidative changes in oocytes and foetal organs we observed here solely to the ageing process. Instead, these alterations likely result from a combination of age‐related OS, metabolic imbalances, and other systemic factors. The harmful effects of both ageing and metabolic dysfunction on reproductive outcomes are well documented. Both ageing and obesity are known to deteriorate oocyte quality and fertility, affecting offspring health as well. Similar or even more pronounced alterations of intracellular lipids and organelles are observed in oocyte and granulosa cells of obese mice (Wu et al. 2010; Jasensky et al. 2016). Studies have shown that maternal obesity leads to the accumulation of saturated fatty acids, ceramides, and other toxic lipid metabolites in oocytes, disrupting lipid homeostasis and impairing oocyte quality (Wu et al. 2015). These lipid alterations are attributed to systemic metabolic dysregulation, including insulin resistance and chronic inflammation, which are hallmarks of obesity (Jungheim et al. 2010). High fat diet also contributes to disrupting oocyte maturation and morphology, accompanied by high levels of ROS (Hou et al. 2016).

These observed oxidative changes in NP AMA F1 and ET AMA F1 prenatal organs indicate the possibility of metabolic homeostasis alterations being established in the developing germ line. Therefore, we next verified the transgenerational persistence of metabolite profile alterations in the adult offspring of AMA mice.

2.3. Abnormal Metabolic Regulation in Offspring and Grand‐Offspring of AMA

There is a wealth of evidence linking gestation at advanced maternal age to impaired metabolic health in offspring, although the mechanisms underlying this association remain unelaborated (Velazquez et al. 2016; Gonzalez et al. 2025). One of the earliest manifestations of this dysfunction is altered weight gain during postnatal development. We observed a significant increase in bodyweight for NP AMA F1 pups at all stages of pup growth, with weights returning to normal by adulthood (Figure 3A). This effect seems to be most seriously felt in the first generation of AMA offspring, as only minor alterations in bodyweight were observed in the second and third generations, suggesting the phenotype is not strongly inherited across generations. However, underlying health consequences may persist, given that we observed a significantly lower survival rate even three generations removed from the initial compromised mating, with NP AMA F3 pups showing lower survival compared to NP YMA (Figure 3B). Increased neonatal mortality is suggestive of impaired developmental resilience or metabolic instability caused by advanced maternal age.

FIGURE 3.

Metabolic effect of AMA persists following embryo transfer to young recipient mothers. (A) Increased bodyweight of NP AMA F1 pups (sample number is given as surviving pups at 5mo); (B) Lower survival rate of NP AMA F3 pups than NP YMA throughout pup development (Gehan‐Breslow‐Wilcoxon test); (C) Decreased bodyweight was observed for ET AMA F3 and ET YMA F3 pups at early stages of development (PND2, PND7) but this did not persist further (sample number is given as surviving pups at 5mo); (D) Normal survival of pups up to weaning; (E) Low serum insulin concentration in NP AMA F2 adult females; (F) FTIR spectra show lower level of saturated fatty acids and cholesterol in NP AMA F1, lower level of phospholipids and sphingolipids in NP AMA F2 and no changes in NP AMA F3; (G) Low serum insulin concentration in ET AMA F3 adult females; (H) FTIR spectra show low level of PI in ET AMA F3. One‐way ANOVA or Kruskal–Wallis test depending on normality, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Interestingly, different effects were observed in the third generation offspring of embryo transfer, both from YMA and AMA‐originating pregnancies. At PND7, both ET AMA F3 and ET YMA F3 pups displayed lower bodyweight than NP YMA pups, while at PND 2 ET YMA F3 pups also had lower bodyweight (Figure 3C). However, this effect was not observed at any other stages of development, so it seems to be a mild alteration. Similarly, no differences in survival were noticed for these groups (Figure 3D).

Given that we observed an increase in bodyweight for NP AMA F1 pups which was not seen in the ET AMA F3 pups, two possible interpretations come to mind: either the transfer to a younger uterus rescues the embryo from this metabolic reprogramming or the bodyweight alterations may have been apparent in the first generation but are less visible by the third. In any case, it is known that embryo transfer surgery (as well as other assisted reproduction techniques) is known to exert metabolic changes on embryos and foetuses, which may lead to bodyweight alterations on offspring (Gardner and Kelley 2017; Garcia‐Dominguez et al. 2020). The stressful conditions during prenatal development caused by non‐physiological conditions which are unavoidable during ET surgery, in combination with AMA, may manifest as Krebs cycle deregulation, lower uptake of metabolites, higher chance of embryo resorption, and epi‐genetic reprogramming (Feil et al. 2006; Ramos‐Ibeas et al. 2019; Garcia‐Dominguez et al. 2020).

Next, we observed differences in serum insulin levels in adult females, with significantly lower insulin in NP AMA F2 and ET AMA F3 individuals compared with NP YMA (Figure 3E,G). FTIR analysis revealed a lower level of circulating metabolites in the first and second generation NP AMA adult offspring, which did not persist in the third generation (Figure 3F). These results point to a broad metabolic reprogramming and a less effective lipid synthesis and transport in AMA offspring mice, which could be explained by increased cellular retention or tissue utilisation of these compounds (Araki et al. 2004).

Interestingly, the male serum displayed a different pattern of metabolic reprogramming than the female, indicating an effect of sex‐linked characteristics (Figure S2). The NP AMA F1 male sera contained a significantly higher insulin level than all other groups (Figure S2A). Higher concentrations of serum insulin were not transmitted to the next generations. Normal insulin concentration was also noted in ET AMA F3 (Figure 1B). Protein secondary structure (Figure S2C,D) was also affected, with a lower α‐helix/β‐sheet ratio for both NP AMA F3 and ET YMA F3 samples compared to NP YMA, indicative of OS conditions. A higher level of metabolites, including PE, PC O‐, NAD+, SAM, SFA, and Chol. in NP AMA F1, F2, and F3 and in ET AMA F3, evidences a persistent effect of serum‐level metabolic reprogramming (Figure S2E,F).

Metabolic deregulation in AMA offspring adults was further evidenced by FTIR analysis of adult liver samples in NP AMA and ET AMA offspring (Figure S3). One of the most consistently deregulated metabolites we observed was phosphatidylethanolamine (PE), as it was found at a lower level in NP AMA F1 and F2 male livers, ET AMA F3 male livers, but at a higher level in NP AMA F1 and ET AMA F3 female livers. PE has several important roles in mammalian cells, including membrane fusion, oxidative phosphorylation regulation, and autophagy (Calzada et al. 2015). It is also a major precursor of phosphatidylcholine (PC), and a dysregulated ratio of PC to PE is known to be a critical symptom of hepatic cellular damage in patients with steatohepatitis (Arendt et al. 2013). Moreover, PE deregulation is indicative of altered redox status, and has been suggested as a biomarker for OS damage (Hisaka and Osawa 2014).

To try to isolate initial causes of this metabolic reprogramming across three generations of AMA offspring mice, we next investigated potential germline mechanisms by examining oocytes from NP AMA F3 mice in comparison with aged controls.

2.4. Metabolic Alterations in AMA‐Derived Oocytes

Considering the links between ageing and elevated OS, and the fact that OS preferentially targets lipids due to their higher susceptibility to peroxidation than DNA or proteins (Ayala et al. 2014), we investigated the lipid profile of germinal vesicle‐stage (GV) oocytes from aged females (NP AMA F0) using FTIR. Interestingly, we found a significant increase in lipids and other metabolites (Figure 4A), represented by elevated absorbance within the FTIR range characteristic for lipids (lipid range: 2800–3000 cm⁻¹) and other chemical compounds (“fingerprint” range: 800–1800 cm⁻¹) (Kreuzer et al. 2020; Böke et al. 2022). This finding spiked our interest to investigate whether these elevated lipid levels in AMA F0 oocytes persist transgenerationally. Remarkably, similar results were also observed in the GV oocytes from young mice of third generation of aged group (NP AMA F3) (Figure 4A,B).

FIGURE 4.

Accumulation of lipids and other metabolites in NP AMA F0 oocytes persists in NP AMA F3. (A) Changes in FTIR spectra in oocytes from NP AMA F0 and NP AMA F3. FTIR range 800–1800 cm⁻¹ refers to fingerprint range and 2800–3000 cm⁻¹ refers to lipids range; (B) FTIR absorbance changes of specific molecular peaks in NP AMA F0 and NP AMA F3 oocytes; mean ± SD. Note that the level of structural lipids (PE, Sph, PS, PC, PI, Hchol, PC O−) in NP AMA F3 remains high or even higher than in NP AMA F0 indicating lipid profile remodelling in NP AMA F3; (C) Hierarchical cluster analysis for FTIR spectra. Left: fingerprint range (800–1800 cm⁻¹), Right: Lipids range (2800–3000 cm⁻¹) showing the affinity of two AMA groups; (D) Principal component analysis of FTIR spectra for fingerprint and lipids ranges show distinct group clusters; (E) Raman spectra changes in oocytes from NP AMA F0 and NP AMA F3. Higher intensity of absorbance noted mainly in fingerprint range; (F) Increased level of oxidative phosphorylation mediators in NP AMA F0 oocytes persists in NP AMA F3; (G) HCA of Raman spectra show the affinity of two AMA groups; (H) PCA for Raman spectra. Data are presented as mean ± SD; one‐way ANOVA, Tukey test, *p < 0.05.

An interesting observation was that NP AMA F0 and NP AMA F3 oocytes present molecular profiles distinct to those from YMA, based on both fingerprint and lipid ranges. Hierarchical clustering (HCA) and principal component analysis (PCA) consistently demonstrate that NP AMA F0 and NP AMA F3 group are closely related, whereas NP YMA oocytes are clearly separated (Figure 4C,D). This shows that lipid alterations in NP AMA oocytes are transmitted across generations despite successive breedings with normal‐aged partners. The persistence of this altered metabolic signature in the F3 generation derived from F2 mothers that were themselves never exposed to an aged uterine environment suggests a mechanism consistent with germline‐mediated inheritance, moving beyond direct intergenerational (F1–F2) effects of the aged maternal milieu.

Complementary to our investigations using FTIR, we also observed a modulated profile of metabolites in NP AMA F3 oocytes with Raman spectroscopy (Figure 4E,F). NP AMA F3 oocytes presented a fingerprint region (800–1800 cm⁻¹) and lipid range (2800–3000 cm⁻¹) more in line with oocytes from aged mothers (NP AMA F0) than with young mothers (NP YMA). HCA and PCA showed a high degree of similarity between NP AMA F3 and NP AMA F0 oocytes in these two ranges of interest, while NP YMA controls formed a distinct cluster (Figure 4G,H). These findings indicate that the metabolic signature of AMA‐derived oocytes is transgenerationally conserved: not only present in the immediate offspring but maintained across multiple generations. Maternal metabolic and environmental conditions can have long‐lasting, transgenerational effects on oocyte composition and offspring health (Aiken et al. 2016). Thus, the persistence of structural lipid changes may be attributed to the fact that during the pregnancy of an aged mother (F0), the developing F1 foetus contains primordial germ cells that will constitute the F2 generation, thereby directly facilitating transgenerational metabolic and oxidative alterations (Yu et al. 2025). Structural lipids in particular are known for their slow turnover rates, which can exceed 2 years in some tissues such as the rat brain (Cuzner et al. 1966).

Next, given the association between OS and mitochondrial phosphorylation, we investigated the regulatory molecular components of OXPHOS in AMA‐derived oocytes. Raman spectroscopy data revealed a significant increase in the key regulators and substrates of mitochondrial OXPHOS, such as cytochrome C, sterol regulatory element binding protein 1 (SREBP‐1), and fatty acids (Figure 4F). An increase in all these components indicates a compensatory metabolic shift to sustain mitochondrial function under stress. SREBP‐1 is a known transcriptional factor of fatty acid synthase, and its expression has been found increased both at the transcriptional and translational levels following exposure to all trans‐retinoic acid (Roder et al. 2007), thus indicating that the retinoid‐responsive regulatory axis can influence lipid biosynthesis and mitochondrial substrate availability.

Complementary to Raman analysis, FTIR spectroscopy data of NP AMA F0 and NP AMA F3 oocytes revealed elevated levels of structural and signalling lipids compared to NP YMA oocytes. These included phosphatidylethanolamine (PE), sphingomyelin (Sph), phosphatidylserine (PS), phosphatidylcholine (PC), phosphatidylinositol (PI), hydroxylated cholesterol (Hchol.), and ether‐linked phosphatidylcholine (PC O−) (Figure 4B). Such enrichment of membrane‐associated lipids suggests an enhanced demand for mitochondrial and cellular membrane remodelling, consistent with increased bioenergetic needs and stress adaptation. This observed accumulation of lipid mediators aligns with prior studies linking maternal metabolic status to oocyte quality across generations. For example, one study demonstrated that a high‐fat/high‐sugar maternal diet leads to transgenerational declines in mitochondrial mass and ATP content in oocytes by the F3 generation, underscoring the heritable impact of metabolic dysregulation (Andreas et al. 2019).

Together, our data suggest that AMA oocytes—including those from the F3 generation—exhibit a heightened demand for OXPHOS components and lipid intermediates. This may reflect an adaptive response to persistent OS, aimed at preserving mitochondrial integrity, supporting energy production, and ensuring oocyte viability.

2.5. AMA‐Derived Oocytes Are Rich in Retinoids That Are Maintained in the Third Generation

Given the importance of Vitamin A components (carotenoids, retinol, retinal and retinoic acid) in OXPHOS function (Acin‐Perez et al. 2010) and based on previous findings demonstrating retinoid localisation in granulosa cells and oocytes, and their association with LD aggregation in aged oocytes (Best et al. 2015; Bisogno et al. 2024), we decided to take a closer look at the retinoid system in AMA‐derived oocytes (Figure 5). To analyse the characteristic vibrations of carotenoids, retinol, retinal, and retinoic acid, spectral deconvolution was applied to both FTIR (Fidge and Goodman 1968) and Raman (Lóránd et al. 2002; Kochan et al. 2015) spectra. We observed increased Raman intensity for peaks corresponding to carotenoids, retinol, retinal, retinoic acid (Kochan et al. 2015) in NP AMA F0 and NP AMA F3 oocytes in comparison to NP YMA oocytes (Figure 5A). These findings were confirmed by FTIR: in NP AMA oocytes we found increased absorbance for retinol, retinal, and retinoic acid in both NP AMA F0 and NP AMA F3 oocytes (Figure 5A,B). While deconvolution significantly enhances spectral interpretation, it is subject to some limitations. The accuracy of the results strongly depends on the initial assumptions regarding the number and shape of component peaks, which may introduce fitting artefacts if not appropriately constrained. Additionally, noise and baseline fluctuations can lead to overfitting or misinterpretation of weak features as distinct bands. To avoid this, care was taken to validate the deconvolution outputs by comparing them with the second derivative of spectra and reference spectra of retinoic acid, retinol, retinal.

FIGURE 5.

Retinoid abundance in NP AMA F0 oocytes persists in NP AMA F3. (A) Left: Raman spectra with highlighted peaks for carotenoids, retinol, retinal, and retinoic acid, Right: corresponding graph shows increase in Raman spectra intensities of carotenoids in NP AMA F0 and F3 oocytes; (B) Left: FTIR spectra with highlighted range corresponding to retinol, retinal, and retinoic acid vibrations; Right: corresponding graph shows increase of carotenoids in NP AMA F0 and F3 oocytes. Mean ± SD; (C) Deconvolution of retinal, retinol, and retinoic acid spectra from Raman (upper row) and FTIR (lower row), show shift of peaks of retinal and retinoic acid in NP AMA F0 and F3; (D) Pearson correlation analysis for Raman and FTIR spectra shows strong positive correlation among retinal, retinol and retinoic acid in NP AMA F0 and F3 oocytes, indicating tightly coupled retinoid metabolism in AMA‐derived oocytes. NP YMA oocytes shows weaker or independent relationships; (E) 2D spectral FTIR correlation analysis shows distinct patterns of retinoid homeostasis for NP YMA (independent), NP AMA F3 (interdependent), and NP AMA F0 (fully coupled) oocytes. Data are presented as mean ± SD; one‐way ANOVA, Tukey test *p < 0.05; Sample size: NP YMA = 10, NP AMA F3 = 20, NP AMA F0 = 19.

2D spectral correlation analysis and PCA analysis of Raman and FTIR data revealed distinct patterns of retinoid regulation across groups (Figure 5D,E). There were weak or independent correlations among retinoid components in NP YMA oocytes, thus suggesting flexible retinoid homeostasis. In contrast, NP AMA F3 oocytes exhibited interdependent signals, while NP AMA F0 oocytes showed strong positive correlations, indicating tightly‐coupled retinoid metabolism in response to aging and accumulated stress.

The conversion of retinol to its fully oxidised form—retinoic acid—via oxidation is an irreversible reaction, and can be interpreted as a sign of heightened OS (Na et al. 2019). In the event of a shortage of retinol inside the oocyte, there is increased transport of retinol and SREBP from the surrounding environment, that is, granulosa cells. Additionally, carotenoids are important precursor molecules which can be used to synthesise retinol, which itself acts as an antioxidant (Lawrence et al. 2004). Given these known facts and the results we describe here, we hypothesise that increased levels of vitamin A components in the oocytes of NP AMA F0 and NP AMA F3 are due to increased OS.

Similarly, in our previous study we found higher levels of carotenoids in the zona pellucida of oocytes from advanced‐age women in comparison to young ones, in addition to increased LD size, area, and number (Bisogno et al. 2024). LD accumulation is a common cellular response to OS, which disturbs lipid metabolism and causes excess fatty acid synthesis (Jarc and Petan 2019). In order to prevent membrane peroxidation, cells sequester these fatty acids by forming new LDs or enlarging existing ones; thus they act as buffers against ROS‐induced lipid damage.

3. Discussion

There are many factors of early life known to cause dysfunctions in later life, as per the Developmental Origin of Health and Disease (DOHaD) hypothesis. These include, but are not limited to maternal nutrition, premature birth, environmental toxin exposure, hormone disruptions, and microbiome changes. But given the globally rising mean age of pregnancy, perhaps none is more relevant to modern times than advanced maternal age. To study the effects of AMA in detail, we analysed oocytes, foetuses, pups, adult tissues, and transgenerational inheritance using a combination of non‐destructive analyses, hormonal assays, redox measurements, and novel detection techniques.

A consistent finding across our datasets is that oocytes from aged females contain elevated lipid levels compared with young controls, consistent with both our prior work and published literature (Gaveglio et al. 2016; Mok et al. 2016; Cordeiro et al. 2018; Musson et al. 2022; Bisogno et al. 2024). To begin addressing this question of why lipids accumulate in ageing female germ cells, we should keep in mind that oocytes are largely dependent on granulosa cells for metabolite supply; they acquire fatty acids from surrounding cells and follicular fluid, storing them in lipid droplets (LDs) for later use. Disruption of this delicate balance, for example absence of supporting cellular neighbours leads to diminished lipid levels (Auclair et al. 2013) or an obesogenic follicular environment, increases lipid transfer into oocytes and embryos (Wu et al. 2010; Leary et al. 2015). Both under‐ and over‐supply of lipids severely impacts oocyte competence and maturation processes. The overloading of fatty acids (i.e., lipotoxicity) we see in AMA oocytes is known to compromise OXPHOS, a key pathway involved in cellular energetics (Ge et al. 2012; Paczkowski et al. 2013, 2014; Li et al. 2020).

Interpretation of lipid accumulation in AMA oocytes is complicated by systemic metabolic changes associated with ageing, including increased BMI (Vuvor and Harrison 2017; Turner and Hubbard‐Turner 2022). Accumulation of lipids in AMA oocytes may therefore simply be a result of metabolic imbalance across the whole body caused by lipid overabundance. After all, many studies have found that metabolic disorders lead to epigenetic reprogramming which can be passed on to offspring. Genes involved in lipid metabolism and insulin signalling are especially affected (Lillycrop et al. 2005; Li et al. 2013). For example, dysregulation of PPARα and PPARγ genes, which regulate fatty acid oxidation and lipid storage, has been observed in offspring of obese mothers (Ornellas et al. 2015; Pang et al. 2021; Guo et al. 2022). Maternal obesity also upregulates SREBP‐1c in offspring, leading to increased lipogenesis and lipid accumulation (Purcell et al. 2023). Offspring of aged or obese mothers are at higher risk of developing obesity, insulin resistance, and cardiovascular diseases later in life (Lansing 1947; Godfrey et al. 2017; Bock et al. 2019; Cochrane et al. 2024).

Another possible explanation may be that the oxidative status of the aged organism directly causes the accumulation of lipids. Notably, while our measurements primarily reflect antioxidant status and metabolite abundance rather than direct oxidative damage, the observed profiles are consistent with a cellular environment experiencing redox imbalance. LDs, beside storing fatty acids, and retinoids can sequester damaged or peroxidised lipids (Henne et al. 2018), protecting membranes like mitochondria and ER—the two organelles most affected in AMA oocytes (van Meer et al. 2008; Liu et al. 2022). It is important to note that peroxidised lipids can decrease membrane thickness and increase permeability to ROS, facilitating further damage to the cell (Yadav et al. 2019). Abnormal accumulation of LDs has been characterised as an early indicator of OS in cancer and neurodegenerative disease in Drosophila and mice (Santos and Schulze 2012; Liu et al. 2015). The increase in lipid levels and increased abundance of LDs we have observed here in AMA oocytes may similarly be interpreted as a marker of oxidative attack and an omen of poor oocyte potential.

Building on this oocyte phenotype, we observed that NP AMA F1 foetuses exhibit impaired development at E16 and E19, evidenced by increased antioxidant levels in foetal liver and brain, reduced placental efficiency, and decreased brain weight (Figure 2). Importantly, ET AMA F1 displayed similar alterations in foetal liver and brain redox status to NP AMA F1, despite gestation in a young uterine environment. These data indicate that the observed oxidative changes are oocyte‐driven rather than uterine‐dependent, highlighting the persistence of AMA effects from the oocyte into foetal development. This interpretation is consistent with existing evidence for mitochondrial dysfunction and metabolic epigenetic alterations in AMA oocytes and blastocysts (Cimadomo et al. 2018; Marshall et al. 2018), as well as adverse effects of AMA on foetal development (Woods et al. 2017). We propose that impaired lipid metabolism originating in the oocyte underlies the observed deficits in foetal organ formation, in line with established links between lipid availability, redox balance, and organogenesis (Chase et al. 1971; van der Kroon and Speijers 1979; Xing et al. 2024).

Decreased foetal‐to‐placental weight ratios and altered placental weight in NP AMA F1 indicate compensatory adaptations to maintain nutrient delivery. However, these adjustments are insufficient to normalise growth, supporting the notion that AMA oocytes program downstream organ impairments. ET AMA F1 foetuses, while maintaining normal placental efficiency and embryo survival, still retain altered redox signatures, further reinforcing the oocyte‐driven origin of oxidative and metabolic perturbations.

The oxidative changes observed in foetal organs in NP AMA F1 and ET AMA F1 manifest later in adulthood as metabolic phenotypes. FTIR analyses of adult liver revealed deregulation of lipid classes, particularly phosphatidylethanolamine (PE) and retinoids (Figure S3). PE is known to be sensitive to peroxidation, and its dysregulation serves as a marker for altered redox status. Retinoids, critical for liver and neuronal function, were also misbalanced, suggesting a broad‐spectrum lipid metabolism shift driven by early‐life oxidative insults. Together, these findings establish a mechanistic trajectory linking AMA oocyte lipid overload to foetal oxidative stress and subsequent adult metabolic reprogramming.

Consistent with previous reports, AMA offspring exhibited bodyweight alterations and metabolic dysregulation (Velazquez et al. 2016), accompanied by serum lipid imbalance, changes in protein secondary structure and disrupted insulin levels. These observations align with the ‘Quiet embryo hypothesis’, which proposes that increased metabolic activity reflects compensatory responses to cellular damage (Aksu et al. 2012; Leese et al. 2022). Sex‐specific differences in serum profiles further highlight the complexity of AMA‐induced programming effects (Figure S2). The elevated neonatal mortality observed across generations (Figure 3) ties advanced maternal age to adverse perinatal outcomes.

Dysregulated lipid metabolism in adult liver and hypothalamus has been implicated in cognitive decline, cellular senescence, and oxidative stress, as observed in conditions such as non‐alcoholic fatty liver disease (Pinçon et al. 2019; Nunes et al. 2022). Altered metabolic phenotypes are closely linked to brain health, and our findings support this association (de Martínez Morentin et al. 2010; González‐Garciá et al. 2017; Hussain et al. 2020). Lipid damage in the brain is particularly problematic due to limited nutrient exchange across the blood–brain barrier and the low turnover of post‐mitotic neurons (Cuzner et al. 1966; de Fabiani 2014; Andreone et al. 2017).

Neurons and oocytes share a vulnerability to lipid damage due to their non‐dividing nature. Lipid alterations in oocytes have been linked to lasting metabolic consequences in offspring, although evidence for intergenerational germline transmission remains limited. Metabolic dysfunctions in oocytes from F1, F2, and F3 offspring of obese mice have been reported (Andreas et al. 2019). In humans, transgenerational transmission of cardiometabolic risk from F0 to F2 generations has been demonstrated (Harville et al. 2017), and high grandmaternal BMI has been associated with increased F2 birthweight, with modifying effects of cigarette smoke exposure (Shen et al. 2020). Other studies have linked ancestral nutrition to offspring longevity (Bygren et al. 2001; Kaati et al. 2007) and to intergenerational risks of obesity and diabetes (Dabelea and Crume 2011). Experimental models further support transgenerational metabolic effects, including altered fecundity and oxidative stress resistance in C. elegans via insulin/IGF signalling (Tauffenberger and Parker 2014) and increased longevity following grandpaternal starvation in Drosophila (Roussou et al. 2016). Collectively, these findings support the potential for lipid alterations associated with AMA or metabolic dysfunction to be transmitted beyond the F1 generation.

Extending these observations, we demonstrate that lipid and retinoid deregulation observed in F1 adult tissues is recapitulated in F3 oocytes, providing direct evidence of transgenerational inheritance of AMA‐induced metabolic and oxidative stress signatures. The retinoid system and its constituent parts are establishing a growing reputation as important mediators of oxidative stress, especially in liver tissues and the prevention of liver diseases (Tsuchiya et al. 2009; Ates‐Alagoz 2013). The retinoid system, in particular, may act protectively in high‐OS environments, consistent with its role in supporting oocyte maturation and mitigating oxidative damage (Maya‐Soriano et al. 2013; Gad et al. 2018; Abdelnour et al. 2019). Together, these results establish a mechanistic link between oocyte lipid/oxidative status, foetal organ redox perturbations, adult organ dysfunction, and transgenerational inheritance.

Our work reports for the first time that high lipid levels in aged oocytes lead to foetal development impairments and oxidative damage to organs, which carry into metabolic reprogramming and lipid balance deregulation in adults and is subsequently transmitted via the female germ line to the third generation (F3). The manifestation of the phenotype in F3, which was not directly exposed to the aged F0 environment, supports its classification as a transgenerational effect potentially mediated by germline inheritance, rather than solely a persistent intergenerational metabolic programming effect, which would most likely be confined to F1‐F2.

4. Conclusion

We present here a holistic overview of the multi‐generational effects of advanced maternal age pregnancies, characterised by OS damage and LD accumulation in oocytes causing developmental impairments in foetuses and metabolic reprogramming in adults, ultimately leading to lipid deregulation in oocytes, thus continuing the transgenerational cycle. In line with the DOHaD hypothesis, we show that conditions at conception and during gestation can exert lasting effects on offspring health, with consequences that may persist across generations.

Our findings contribute to a growing body of evidence emphasising that assisted reproductive technologies (ART) should move beyond the assessment of gross morphological features and instead incorporate detection of subtler cellular and molecular alterations in germ cells. The introduction of new methods to the market will enable quality assessment of oocytes before fertilisation.

Emerging advances in non‐destructive damage detection, artificial intelligence–based predictive modelling, precise robotic embryo handling, and personalised treatment protocols offer promising opportunities to refine oocyte selection and improve long‐term outcomes for offspring. Early applications of these approaches are already beginning to demonstrate their potential (Mapari et al. 2024), positioning ART at the forefront of a rapidly evolving field with transformative implications for reproductive medicine.

5. Methods

All chemicals, unless otherwise indicated, were obtained from Sigma‐Aldrich.

5.1. Animal Studies and Embryo Transfer

Animal experiments were performed on 8–10 months old (AMA) and 3–5 months old (YMA) mice. C57BL/6 mice (n = 85) were used. Animals were maintained in temperature‐ and light‐controlled conditions (22°C and 12 h light–dark cycle) and were provided with food and water ad libitum. All experimental procedures were conducted according to the guidelines of European Community Regulations and conformed to the Polish Governmental Act for Animal Care. Animal procedures conducted were approved by the II Local Ethical Commission of Kraków (123/2018; 318/2018; and 329/2021). AMA and YMA females were paired with stud males overnight, and vaginal plugs were checked the following morning to confirm mating. Part of plug positive females were euthanised by cervical dislocation 4 days post‐coitum (dpc) to collect embryos at blastocyst stage. Embryos were recovered by flushing each uterine horn with 1 mL of M2 medium and transferred to 3 month‐old CBAxC57BL/6 pseudo‐pregnant females (ET AMA and ET YMA groups), as described previously (Arena et al. 2021). Other plug‐positive females were allowed to continue the pregnancies to term (natural pregnancy groups: NP AMA and NP YMA). Some pregnant females in groups NP YMA, NP AMA, ET YMA, and ET AMA were ethically sacrificed either at 16 or 19 dpc to collect foetal organs for analysis.

5.2. Transgenerational Study and Tissue Collection

To study the transgenerational inheritance of AMA‐derived effects, we continued the mating of NP AMA F1, ET YMA F1, and ET AMA F1 until the third generation (NP AMA F3; ET YMA F3, ET AMA F3). 3–5‐month‐old male and female AMA offspring were mated with age‐matched YMA animals. Male and female F1, F2, and F3 mice were sacrificed for organ and oocyte collection at 5 months of age. Animals were euthanised by decapitation. Trunk blood was collected and allowed to rest on ice until coagulation, then centrifuged at 12,000g for 15 min at 4°C to obtain serum fraction. Other organs, including liver and brain sections, were collected and stored at −80°C until analysis.

5.3. Collection of GV Oocytes

GV stage oocytes were collected from ovaries of sacrificed females and treated with 100 IU/mL of hyaluronidase to remove surrounding cumulus. Oocytes then underwent fixation process by washing in PBS + 0.4% PVP three times followed by incubation in 4% PFA for 30 min and washing again in PBS + 0.4% PVP before proceeding analysis.

5.4. Coherent Anti‐Stokes Raman Spectroscopy

LDs distribution in MII oocytes was evaluated using CARS as described previously (Arena et al. 2021). The multimodal nonlinear microscope used in this study consisted of a Leica DMi8 inverted microscope equipped with the Leica TCS SP8 CARS module and the Leica SP8 confocal module (Leica Microsystems). The Leica TCS SP8 CARS uses a tuneable pump laser with a tuning range of 780–940 nm, combined with a Stokes laser at 1064 nm, provided by Laser picoEmerald and integrated with a 750‐mW power optical parametric oscillator. The oocytes were placed on a Lab‐Tek chamber (Thermo Scientific). To detect the LD signal, the optical parametric oscillator's wavelength was tuned to 816.7 nm to serve as the pump beam, in combination with the 1064‐nm Stokes beam to probe the CH₂ stretch vibration. All signals were detected using a photomultiplier tube by collecting the photons in a forward direction through a 40× objective. Each oocyte was scanned across all of its sizes in the z axis using a z step of 1.5 μm. For the analysis, all stacks from a single sample were processed to obtain a maximum projection image. The threshold was adjusted specifically for all samples to collect all positive signals. When necessary, LDs were separated manually in cases of obvious pixel overlay. Statistics regarding the LDs area and number were generated automatically by LAS X software (Leica Microsystems). The LAS X software was used to apply colour‐coded LD size masks to the maximum projection images for graphical observation of the LDs' size distribution of the analysed samples. The oocyte area was calculated using ImageJ (NIH) due to limitations of the LAS X software.

5.5. Foetuses and Placentae Collection

Pregnant mice were ethically sacrificed by cervical dislocation. A midline incision was made through the abdominal wall muscles to expose the internal organs. Uteri with prenatal pups were located and carefully removed. After removing the uterine tissue and yolk sac, placentae and foetuses were weighed, followed by organ collection from foetuses after ethically sacrificing pups. Foetal brain and liver were stored at −80°C until analysis.

5.6. GSH and GSSG Activity Assays

Reduced (GSH) and oxidised (GSSG) glutathione levels in foetal liver and brain tissues were quantified using the Glutathione Colorimetric Detection Kit (Invitrogen, Cat. No. EIAGSHC), following the manufacturer's protocol. Absorbance was measured at 405 nm, and concentrations of GSH, GSSG, and total glutathione were calculated according to the kit's standard curve and were plotted after normalisation with protein content of tissue lysate.

5.7. Insulin Assay

Total insulin concentration in mouse serum was determined via use of the Rat/Mouse Insulin ELISA Kit (Cat No. EZRMI‐13K; Merck KGaA, Darmstadt, Germany), following the manufacturer's protocol. Absorbance was measured at 450 nm minus a background correction at 590 nm and quantified using a nonlinear regression standard curve generated from kit‐provided standards.

5.8. Raman Spectroscopy

Raman Spectroscopy was performed as we previously described (Bisogno et al. 2024). Briefly, Raman spectra of mouse oocytes were collected using the MonoVista CRS3 500 system. For this purpose, oocytes were placed on the CaF₂ Raman grade slides and measured using a 10× objective lens and a 532 nm laser, with the laser power calibrated for optimal signal‐to‐noise ratio. The exposure time was set to 10 s per accumulation. The resulting Raman spectra were processed using OPUS software, which involved baseline correction and min‐max normalisation.

5.9. Fourier Transform Infrared (FTIR) Spectroscopy

FTIR was performed as we previously described (Bisogno et al. 2024). Briefly, FTIR spectra of mouse oocytes were collected using a Nicolet iS50 FTIR spectrometer (Thermo Scientific) with the attenuated total reflectance technique. To obtain FTIR spectra, 2 μL of PBS solution containing a mouse oocyte was dropped onto the CaF₂ crystal and left to sit for 5 min until the PBS evaporated. Before oocyte measurements, the background spectrum was collected using the same parameters as for oocytes: 32 scans, 4 cm⁻¹ spectral resolution, range from 400 to 4000 cm⁻¹. After measurements, for all spectra, baseline correction and min‐max normalisation were performed. Measurements as well as the analysis of the obtained spectra were performed by the OPUS 7.0 software. For FTIR analysis of serum, 4 μL of thawed sample was directly pipetted onto a CaF₂ slide, and measurements were made as normal. For hypothalamus and liver analysis, a small piece of thawed tissue of equal mass per sample was manually homogenised in 50 μL PBS with disposable polypropylene pestles (SP Bel‐Art); then 4 μL was added to CaF₂ slides and measured when dry.

5.10. Statistical Analysis

One‐way ANOVA with Tukey post hoc test was performed for determination of statistical significance in the Raman intensity or absorbance value of bands building phospholipids, carotenoids, proteins, and lipid structures in three analysed groups of samples. Chemometric analyses such as Principal Component Analysis (PCA) and Hierarchical Cluster Analysis (HCA) were performed for two fingerprint ranges (800–1800 cm⁻¹, (Böke et al. 2022)–FTIR; (Wu et al. 2024)–Raman) and lipids wavenumbers (2800–3000 cm⁻¹, (Derenne et al. 2014)–FTIR; (Czamara et al. 2015)–Raman). Respectively, 2076 and 417 points from each sample were used. First analysis was done to show the possibility of distinguishing two analysed groups using Raman and FTIR spectroscopy, and second to show similarity between samples from the same or from other groups. In PCA, 95% of significance were marked, and in HCA paired group (UPGMA) algorithm and Euclidean similarity index were used.

5.11. Deconvolution of Raman and FTIR Spectra

Clear discrimination of overlapping vibrational bands in complex molecular systems can obscure subtle spectral features, especially in regions where multiple compounds contribute to closely spaced peaks. The deconvolution process allowed for the resolution of individual bands within the congested spectral regions by mathematically modelling the observed spectrum as a sum of narrower, theoretically plausible component peaks. These were fitted using appropriate peak profiles with their positions, widths, and intensities optimised to best match the experimental data. The fitting procedure was guided by known vibrational modes for the spectra of relevant compounds (Fidge and Goodman 1968; Lóránd et al. 2002; Kochan et al. 2015), ensuring that the resolved peaks were chemically meaningful. Deconvolution was performed using MagicPlotStudent software.

Author Contributions

Hafsa Gulzar, Richard Musson performed research, analysed data and wrote the paper. Simona Bisogno, Lulu Alluhaibi, Alexey Maximenko, Ewelina Bik, Małgorzata Barańska performed research. Joanna Depciuch performed research and analysed data. Grażyna E. Ptak designed and supervised research and revised the final version of the paper.

Funding

This work was supported by Narodowe Centrum Nauki, 2021/41/B/NZ3/03507, 2019/35/B/NZ4/03547.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Aged females have increased bodyweight. High bodyweight of NP AMA F0 mice across different mice strains. Mean ± SEM, Two‐tailed t‐test, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure S2: Increased insulin resistance and changes in serum compounds in adult male offspring of aged mothers and grandmothers. (A) Higher insulin concentration in NP AMA F1 serum; (B) Normal mean insulin concentration in ET AMA F3; (C) Lower α‐helix/β‐sheet ratio in NP AMA F3; (D) Lower α‐helix/β‐sheet ratio in ET YMA F3. ***p < 0.001, ****p < 0.0001; (E) FTIR spectra show higher level of PE, PC O−. NAD⁺, SFA and Chol in NP AMA F1; higher level of NAD⁺, SFA, Chol in NP AMA F2 and higher level of glucose, SFA and Chol in NP AMA F3 serum; (F) FTIR spectra showed lower level of PE, PC O−, Glu, SAM, NAD⁺, AEA, Chol. and PUFA in ET AMA F3 versus NP YMA and lower level of PE, SFA, Chol. and PUFA versus ET YMA F3; PCA shows differentiation between analysed groups.

Figure S3: Male and female liver FTIR analysis. (A) Male NP—Upper row: FTIR spectra showed lower level of PE and retinol in NP AMA F2, lower level of PE, retinoic acid, retinal and retinol in NP AMA F1, Lower row: PCA showed differentiation between NP AMA F1 and NP YMA, between NP AMA F2 and NP YMA, and between NP AMA F1 and NP AMA F2; changes in the protein structure in NP AMA F1 and NP AMA F2 compared with NP AMA F3; (B) Male ET—FTIR spectra showed lower level of retinal, retinol and PE in ET AMA F3 versus NP YMA and lower level of retinol and PE in ET YMA F3 versus ET YMA F3; PCA showed differentiation between ET AMA F3 and NP YMA group; no changes in membrane fluidity but several changes in the secondary structure in proteins in ET AMA F3; (C) Female NP—Upper row: FTIR spectra showed higher level of PE in NP AMA F1, Lower row: PCA showed differentiation between all analysed groups; lower membrane fluidity in NP AMA F1 than NP YMA, no changes in the protein structure in NP AMA F1 in compared with NP YMA; (D) Female ET—FTIR spectra showed higher level of retinol and PE in ET AMA F3, as well as in ET YMA F3 and no changes in ET AMA F3 versus ET YMA F3; PCA showed differentiation between ET AMA F3 and NP YMA group; no changes in membrane fluidity and no changes in the secondary structure of proteins in ET AMA F3. One‐way ANOVA or Kruskal–Wallis test depending on normality; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.