Alzheimer’s disease (AD) is a progressive neurodegenerative disorder marked by impaired behavior, cognitive dysfunction, and episodic memory loss; being the most common dementia in the elderly and the most prevalent neurodegenerative disease (Masters et al., 2015). From the genetic standpoint, AD can be divided into two different categories: the familiar form (<10% of all cases) and the sporadic form (90–95% of cases). The familial form (FAD) has an early onset. It is triggered by mutations in any of the three primary genes linked to AD: APP (amyloid precursor protein) gene, PSEN1 (presenilin) or PSEN2 (presenilin) gene (Dorszewska et al., 2016) in chromosomes 21 and 14, respectively. These genes contribute to the amyloidogenic pathway by which the APP protein is consecutively cleaved and processed to produce Aβ oligomers susceptible to aggregate (Hardy and Selkoe, 2002). On the other hand, sporadic AD (SAD) results from a complex mixture of genetic and environmental factors. However, its pathobiology is still under investigation (Dorszewska et al., 2016).

The allele ε4 of the APOE gene encoding the apolipoprotein E (APOE) is the major genetic risk factor for AD (Fernández-Calle et al., 2022). Within the central nervous system, astrocytes mainly produce this protein. It is critical in shuttling cholesterol to neurons to maintain cell membranes and synapses and allow their reparation after injury (Belloy et al., 2019; García-Sanz et al., 2021). In humans, there are three APOE isoforms, which differ only by 1 or 2 amino acids: APOE2, APOE3, and APOE4. The ε3 allele has the highest prevalence (78%), followed by ε4 (15%) and finally ε2 (7%) (Masters et al., 2015). The ε4 allele in homozygosis raises the risk of developing AD by twelve times (Heffernan et al., 2016). The ε4 allele is linked to an earlier onset of the disease (Khachaturian et al., 2004) and more severe cognitive impairment (Najm et al., 2019). It has an approximately 50% contribution to the development of SAD (Ashford, 2004). Furthermore, the presence of the ε4 allele is related to the appearance of amyloid-β (Ab) aggregates, the hyperphosphorylation of tau, and the disorganization of mitochondrial networks (Mahley et al., 2007; Cheng and Bai, 2018) and ultimately, it may account for specific phenotypic heterogeneity in AD (Emrani et al., 2020).

The senile plaques composed of extracellular Aβ aggregates and neurofibrillary tangles formed by hyperphosphorylated tau protein remain AD’s two most critical histopathological hallmarks (Lane et al., 2018; Vidal and Zhang, 2021). Growing evidence suggests that abnormal mitochondrial function is involved in AD pathophysiology (Castora et al., 2022). Environmental toxins, such as pesticides, heavy metals, and industrial waste products, can impair mitochondrial function and produce reactive oxygen species (ROS) and oxidative stress, which can damage neurons and contribute to AD pathogenesis (Sharma et al., 2021). Mitochondrial dysfunction is observed in AD subjects within the brain and systemically (Strope and Wilkins, 2023). AD is depicted by disrupted energy metabolism in the brain and increased levels of oxidative stress (Wang et al., 2020). Furthermore, Aβ aggregates may also destabilize Ca²⁺ homeostasis generating a Ca²⁺ overload in the mitochondria and forming a permeability transition pore in the inner mitochondrial membrane (Wacquier et al., 2020). Eventually, this event may trigger the release of cytochrome C from mitochondria and the collapse of the mitochondrial membrane potential (Brookes et al., 2004; Calvo-Rodriguez and Bacskai, 2021). Moreover, Ca²⁺ can increase the number of reactive oxygen species (ROS) through two mechanisms: nitric oxide production, which inhibits the mitochondrial IV complex, and the increase in the activity of the electron transport chain, which results in increased ROS production (Brookes et al., 2004; Calvo-Rodriguez and Bacskai, 2021). Indeed, defective mitophagy mediated a preserved mechanism of memory loss across the AD models (Xie et al., 2022; Zeng et al., 2022). The impairment of the endocytic, autophagic, and lysosomal pathways is considered initiated at the early stages of AD. It appears to be involved in most AD cases (Krance et al., 2022) and other neurodegenerative diseases (García-Sanz et al., 2018). Autophagy appears to be compromised in not only AD but also in other neurodegenerative disorders. Autophagy is essential in maintaining normal cell function by removing potentially harmful materials, including damaged organelles (such as mitochondria) and poorly folded or aggregated proteins (Vegh et al., 2019; Zeng et al., 2022). Defects in this mechanism lead to an accumulation of these toxic materials, eventually causing neuronal death as it occurs in some neurodegenerative diseases, including AD (Filippone et al., 2022; Griffey and Yamamoto, 2022).

The initiation of the autophagy pathway is regulated by a plethora of different proteins, the most remarkable of which are the mTOR and ULK complexes. In the presence of nutrients, mTOR is phosphorylated and inhibits autophagy through the phosphorylation (at specific inhibition sites) of ULK, which is one of the proteins that contribute to the initiation of the phagophore formation process. On the contrary, in nutrient deprivation situations, mTOR will stop inhibiting ULK so that phagophore formation can occur and, therefore, autophagy will be enhanced (Alers et al., 2012). In AD, mTOR and other proteins involved in autophagosome formation are especially susceptible to changes in their function due to oxidative stress (Lee et al., 2012; Filippone et al., 2022). Altered autophagy is broadly established in AD, leading to damaged organelles buildup, including mitochondria (Frake et al., 2015). However, current research remains controversial regarding which stages of autophagy are specifically impaired. A comprehensive assessment of the autophagic process in CA1 pyramidal hippocampal neurons from early and late-stage AD patients showed a remarkable upregulation of autophagy-related genes, reflecting increases in both autophagosome and lysosome biogenesis. This induced autophagy status appears to be an early mechanism response, and autophagy flux is gradually hampered due to the failure of the lysosomal degradation (Bordi et al., 2016).

In fact, with age and under stress, lysosomes accumulate lipofuscin, which cannot be degraded, leading to lysosomal dysfunction (Brunk and Terman, 2002; Trigo et al., 2022). Strikingly, in various lysosomal storage diseases, lysosomal defects initially produce a burden of amyloidogenic proteins (Monaco and Fraldi, 2020; Riera-Tur et al., 2022). In AD, allele ε4 of the APOE gene has been related to alterations in the endocytic, autophagic, and lysosomal processes (Schmukler et al., 2018; Eran and Ronit, 2022; Fernández-Calle et al., 2022). In addition, it has been shown that PSEN1 mutations produce lysosomal and autophagic dysfunction due to defects in lysosomal acidification and lysosomal Ca²⁺ homeostasis (Coen et al., 2012; Lee et al., 2015; Yang et al., 2019; McDaid et al., 2020).

Overall, AD pathology generates increased oxidative stress and alterations in autophagy that compromise cellular homeostasis, favoring the mechanisms of neurodegeneration. In this study, we evaluate the impact of the presence of the APOEε4 allele and a mutation in PSEN1 over cellular viability, ROS production, mitochondrial structure, autophagy status, and lysosome accumulation and distribution in age-matched fibroblasts from healthy and AD patients. We used fibroblasts because they are easily isolated from skin biopsies preserving the chronological and biological aging of patients and their environment etiopathology. Indeed, they are extensively used as the model of several neurodegenerative disorders, including AD (Pani et al., 2009; Theendakara et al., 2016; Pérez et al., 2017; Olesen et al., 2022).

In this study, we evaluated the impact of the ε4 allele of the APOE gene and the mutation G206D in PSEN1 on the molecular mechanisms leading to AD pathology. We found that PSEN1 and APOE4/4 or APOE3/4 confer different phenotypes in human fibroblasts, similarly to that reported in iPSC-derived human microglia; however, in microglia, PSEN1 and APOE4/4 or APOE3/4 affects other underlying mechanisms (Konttinen et al., 2019). Our results show alterations in the vulnerability of AD fibroblasts to oxidative stress and a disruption in the mitochondrial network of APOE 3/4 + PSEN1 and APOE 4/4 fibroblasts, as summaries in Figure 7. Moreover, the mutation in PSEN1 also affects the autophagy pathway and the lysosomal function, increasing lysosomal accumulation and clustering along with an increase of global p62/SQSTM1 (Figure 7).

First, we found that APOE 4/4 and APOE 3/3 genotypes confer protection to human fibroblasts against oxidative stress-induced cellular vulnerability (Figures 1A–D), which disagrees with the increased ROS described in other AD studies (Pérez et al., 2017; Sarasija et al., 2018; Drabik et al., 2021). This is particularly striking in the case of the APOE 4/4 genotype because it contributes to mitochondrial respiratory chain disruption (Orr et al., 2019) and impairs mitochondrial neuron function in vivo and in vitro (Liang et al., 2021). In this regard, a study detected an increase in ROS in the plasma of APOE 4/4 AD patients (Massaccesi et al., 2019). However, in this case, the differences observed with our results could be due to blood analyses not exhibiting accurate intracellular or tissue ROS content. Moreover, one of the natural mechanisms by which the cell tries to protect itself against oxidative damage is the arrest of the cell cycle to enable the repair of ROS-induced DNA damage. This occurs through the activation of p53, which triggers cell cycle arrest, DNA repair, and activation of apoptosis (Szybińska and Leśniakx, 2017). Fibroblasts from AD patients exhibited ROS-mediated p53 activation, implying that these cells are less susceptible to oxidative stress and, thus, more resistant than control fibroblasts (Naderi et al., 2006; Uberti et al., 2006). This could be why APOE 4/4 and APOE 3/3 fibroblasts are less prone to oxidative stress-induced cell death. Moreover, p53-induced antioxidant response in vivo could develop a neuroprotective function. However, the neuroprotective effects of p53 in AD remain controversial (Abate et al., 2020).

APOE is an apolipoprotein with an essential function in cholesterol trafficking. The presence of the APOE ε4 allele has typically been related to a lipid homeostasis imbalance in AD patients since this isoform has lower transport affinity and binding capacity for lipids and, in particular, cholesterol (Chang et al., 2017; Jeong et al., 2019; Lanfranco et al., 2020). Previous studies with Filipin staining show that AD fibroblasts (Pani et al., 2009) and, more specifically, APOE 4/4 astrocytes (Lin et al., 2018) present higher levels of free cholesterol compared to controls. Contrary to these, our results do not show evidence of this accumulation (Figures 2A, B). However, we did surprisingly find a remarkable reduction of the free cholesterol levels inside APOE 3/4 + PSEN1 fibroblasts (Figure 2B). This could be due to a defect in its synthesis, which can eventually lead to a disruption in organelle membranes and cell death (King et al., 2016). Therefore, our results could indicate that the observed decreased cell viability of APOE 3/4 + PSEN1 fibroblasts could probably be enhanced by this abnormal cholesterol depletion. We also quantified the free cholesterol levels inside lysosomes in our fibroblasts (Figure 2C). Although no significant results were obtained, we observed a slight increment of free cholesterol in APOE 3/3 and APOE 4/4 fibroblasts and a slight lessening in APOE 3/4 + PSEN1 fibroblasts. Other authors described that the lysosomal accumulation of cholesterol (García-Sanz et al., 2021) could rescue cells from lysosome-dependent cell death (Appelqvist et al., 2011; King et al., 2016). Hence, this lysosomal cholesterol build-up could be one of the reasons why these APOE 3/3 and APOE 4/4 fibroblasts appear to be protected against cell death. Defects related to lysosome dysregulation, lipid membrane disruption, intracellular cholesterol distribution, and altered Ca²⁺ signaling depend on the APOEε4 allele and sex in immortalized astrocytes (Larramona-Arcas et al., 2020). Therefore, this could be a possible explanation for our APOE 4/4 results.

Aging stands out as the most pivotal risk factor for neurodegenerative disorders, including AD (Hou et al., 2019). Aging decreases the cellular ability to produce energy (Ozgen et al., 2022; Trigo et al., 2022), and mitochondria play an essential role in producing such energy. They are necessary for regulating critical biochemical processes such as Ca²⁺ storage and homeostasis, activation of the oxidative stress response, and cell death pathways (Ribas et al., 2014). Therefore, mitochondrial dysfunction is closely linked to AD pathogenesis (Swerdlow, 2018; Perez Ortiz and Swerdlow, 2019; Castora et al., 2022). In this context, our results showed that fibroblasts from patients with AD have mitochondria with fewer reticular networks (Figures 3D–I, 5). This is especially evident in APOE 3/4 + PSEN1 fibroblasts, which present dot-shaped mitochondria (Figure 3D). This dotted appearance is observed in other studies as an indicator of fragmented mitochondria in unhealthy and oxidatively stressed cells (Ježek and Plecitá-Hlavatá, 2009).

Moreover, AD fibroblasts, especially those with the ε4 allele of the APOE gene in homozygosis, have higher mitochondria labeled with p62/SQSTM1 (Figure 4A), a fragmented mitochondrial network (Figures 3A–E). This could be due to the early enhanced induction of autophagy in AD fibroblasts, as previously described (Bordi et al., 2016). This induction would eventually lead to a higher degradation of damaged mitochondria. In this regard, a slight disruption of the mitochondrial network was detected in all our fibroblasts from AD patients (Figures 3D–I). From the mitochondrial perspective, this could entail a common pathogenic origin of the disease, regardless of the genotype (Yin et al., 2020). Specifically, APOE 4/4 fibroblasts exhibited increased mitochondrial fragmentation as indicated by the significant increase of individual fragments (Figure 3E) as was shown in other studies (Cabezas-Opazo et al., 2015; Pérez et al., 2017). It was recently determined that despite also showing an impaired mitochondrial network, APOE4 astrocytes displayed an increased number of branches and fewer individual mitochondria, contrary to our findings (Schmukler et al., 2020). Impairments in the mitochondria of these astrocytes are also supported by deficiencies in their synthesis, recruitment, ubiquitination, fission, fusion, and mitophagy (Eran and Ronit, 2022). The mitochondrial disorganization we have found is more evident in the case of APOE 3/4 + PSEN1 fibroblasts due to the decreased size of the mitochondrial network, the length of its branches, and the area occupied by mitochondria (Figures 3G–I). This latter parameter, which may also be indicative of the percentage of mitochondrial biomass in the fibroblasts (Sinha et al., 2019; Li et al., 2020), suggests that APOE 3/4 + PSEN1 fibroblasts could have a reduction in the overall mitochondrial mass. These mitochondrial abnormalities found in APOE 3/4 + PSEN1 fibroblasts agree with those previously found in fibroblasts from AD patients with PSEN1 mutation (Gray and Quinn, 2015; Bell et al., 2018). This impaired integrity of the mitochondrial network could be responsible for the subtle higher superoxide anion levels and increased vulnerability to oxidative stress of APOE 3/4 + PSEN1 fibroblasts. This mechanism may be primarily caused by the deregulation of the Ca²⁺ homeostasis induced by the mutation in PSEN1, as described in PSEN1 mutant AD iPSC-derived astrocytes (Oksanen et al., 2017). Mitochondria form an interconnected network, which allows them to communicate rapidly and distribute energy throughout the cell (Trigo et al., 2022). However, this connectivity puts the energy conversion system at risk because the entire network could suffer the consequences if any elements are damaged (Trushina et al., 2012). Mitochondria are dynamic organelles constantly fusing and dividing (Bertholet et al., 2016). The dynamic equilibrium between fusion and fission phenomena defines the morphology of the mitochondria, allowing their adaptation to energy needs. Therefore, the increased mitochondrial fragmentation of AD fibroblasts suggests a possible fusion and fission balance disruption. Moreover, previous studies on AD also demonstrated that oxidative stress increases the fragmentation of the mitochondrial network via the deregulation of mitochondrial fusion and fission dynamics (Zhu et al., 2012; Misrani et al., 2021; Olesen et al., 2022).

Another critical mechanism in controlling mitochondrial quality is mitophagy. The defects in mitochondrial networks’ complexity in APOE 4/4 and APOE 3/4 + PSEN1 fibroblasts (Figure 3), aligns with the significantly higher levels of p62/SQSTM1 detected in these fibroblasts (Figure 4B), which probably indicates an induced mitophagy to eliminate these harmful organelles (Figure 7). Consistently, APOE 4/4 fibroblasts also had a higher percentage of mitochondria labeled with p62/SQSTM1 destined for degradation, as shown by the colocalization analysis (Figure 4C). However, fibroblasts with the PSEN1 mutation did not show an increased colocalization of their mitochondria with p62/SQSTM1 compared to controls (Figure 4C). These results could be explained because of a defect in mitochondrial biogenesis processes, which would result in aberrant mitochondria, thus, contributing to the pathophysiology of AD, as already demonstrated in previous studies where mitochondrial biogenesis markers such as PGC-1a are reduced (Sheng et al., 2012). Another possibility is that APOE 3/4 + PSEN1 fibroblasts undergo an ineffective mitophagy process. This would align with previous studies in PSEN1 fibroblasts and iPSC-derived neurons from patients, where mitochondria were labeled correctly but unable to be degraded (Trushina et al., 2012). In addition, APOE 3/3 fibroblasts do not present significant changes in mitochondrial biomass (Figure 3), p62/SQSTM1 levels, and mitochondria labeled with p62/SQSTM1 compared to controls (Figure 4). These results align with the ε3 allele of the APOE gene being less related to the pathophysiology of the disease (Heffernan et al., 2016), as also demonstrated in mouse model studies (Simonovitch et al., 2019).

A dampened lysosomal autophagic clearance or an altered activation of autophagy could be responsible for the aforementioned increased p62/SQSTM1 in APOE 3/4 + PSEN1 and APOE 4/4 fibroblasts (Figure 7). Similarly, higher levels of p62/SQTM1 and parkin were found in the hippocampus of APOE4 mice compared to APOE3 mice, showing reduced mitophagy (Simonovitch et al., 2019). In addition, even though the analysis of the levels of p-mTOR showed no conclusive outcome (Figures 5A, B), in the case of ULK, there was a slight decrease of p-ULK in all AD fibroblasts compared to controls (Figures 5A, C). Beclin1 was also significantly increased in APOE 3/3 and APOE 3/4 + PSEN1 fibroblasts (Figures 5D, E). This is consistent with the increased colocalization of mitochondria with p62/SQSTM1 detected in APOE 4/4 fibroblasts (Figure 4C), indicating a possible impaired mitophagy. Therefore, these results could indicate that fibroblasts from AD patients may develop an altered and induced autophagic pathway onset, which differs from previous studies with AD patients that show an accumulation of defective mitochondria due to defects in autophagy induction (Martín-Maestro et al., 2017b; Vegh et al., 2019). This could suggest that PSEN1 mutation might not only accelerate autophagosome synthesis through Beclin1 and mTOR pathways but also alter autophagic clearance (increased p62/SQSTM1) likely due to a secondary degradation defect in lysosomes, triggering an autophagosome buildup (LC3II accumulation after CQ treatment). However, as mentioned before, it cannot rule out that this effect could be due to the non-canonical autophagy activation (Jacquin et al., 2017; Fletcher et al., 2018). In either case, the results could point to an anomalously induced autophagy, probably as a compensatory mechanism due to a decline in the degradation of the last step of autophagy (Martín-Maestro et al., 2017a). Moreover, it must be considered that there is still debate about whether autophagy is altered in single or multiple stages in AD (Bordi et al., 2016), and further studies regarding this are still needed to clarify this question.

Consistently with the possible degradation failure, we found lysosomal impairment in AD fibroblasts. We found higher levels of LAMP1 in APOE 3/4 + PSEN1 fibroblasts (Figures 6A, B) which aligns with previous studies displaying elevated LAMP1 levels and its mRNA in AD patients’ cortexes (Barrachina et al., 2006). PSEN1 mutations generate a defect in the N-glycosylation of the V0a1 subunit of the v-ATPase, causing problems in its transport toward the lysosomes. This fact leads to defects in the acidification of the lysosomes, as well as deficiencies in their proteolysis (Lee et al., 2010; Vegh et al., 2019). Also, under acute mitochondrial stress conditions, AMP-dependent protein kinase (AMPK) is repressed, leading to an accumulation of lysosomal Ca²⁺ and a loss of lysosomal hydrolysis due to defects in acidification (Deus et al., 2020). Thus, the accumulation of lysosomes in our PSEN1 fibroblasts could be due to these acidification deficiencies. This would result in dysfunctional lysosomes that tend to accumulate, increasing their clustering rate (Figure 6C), as previously described in APP and PSEN1 mutant neurons (Hung and Livesey, 2018).

Furthermore, the higher proportion of perinuclear lysosome clusters in all our AD fibroblasts (Figures 6D, E), even more, enhanced in those with the APOE 3/4 + PSEN1 genotype could also be an indicator of lysosomal disturbance, probably by blocking the lysosomal exocytosis. The cellular distribution of lysosomes is relevant in modulating lysosomal function and coordinating cellular responses to the presence or absence of nutrients (Tancini et al., 2020). One of the cellular responses coordinated by the changes in the intracellular localization of lysosomes is the process of autophagy since mTOR, the primary regulator of this pathway, is found inside lysosomes. The position of lysosomes within the cell changes in response to nutrient availability, thus coordinating the mTOR activity and the successive autophagy induction. When nutrients are scarce, there is an increase in the intracellular pH, and lysosomes move toward the perinuclear region. This causes the inactivation of mTOR, which activates autophagy, facilitating the fusion of the autophagosome with the lysosome. Conversely, when nutrients are available, cytoplasmic pH decreases, lysosomes return to peripheral regions, and mTOR is activated again, thus inhibiting autophagy (Korolchuk et al., 2011). Consequently, AD-derived alterations in this lysosomal transport could be responsible for this perinuclear clustering phenotype, as already described (Kanaan et al., 2013; Hwang et al., 2019; Lie and Nixon, 2019).

In brief, in this study, we have assessed the impact of the ε4 allele of the APOE gene and a mutation in PSEN1(G206D) on the cellular mechanisms underlying the pathogenesis of AD using skin fibroblasts derived from AD patients (Figure 7). Although this cellular model has some limitations, this experimental approach has allowed us to obtain significant differences between control and patient-derived fibroblasts in several parameters. Hence, fibroblasts can be a good model for studying pathological mechanisms in AD, since they constitute an easily accessible patient-specific cellular model of the disease. This is due to the cellular plasticity of skin fibroblasts, which endows them with the potential to be easily cultured and have levels of gene expression and damage accumulation similar to those of neurons (Bell et al., 2018; Tong et al., 2022).

Alzheimer’s disease, as in other neurodegenerative diseases, has a systemic element that can affect peripheral cells outside the nervous system, characterized by a series of changes at the metabolic level, such as alterations in autophagy or mitochondrial dysfunction. Therefore, the study of these changes in fibroblasts derived from AD patients can contribute to the deciphering of the molecular physiopathology of the disease. We found that the APOE allele ε4 in homozygosis produces an increased fragmentation of the mitochondrial network, probably due to slightly induced mitophagy to eliminate these damaged mitochondria. Moreover, PSEN1 mutation disrupts the integrity of the mitochondrial network, triggering high superoxide anion levels and, thus, making APOE 3/4 + PSEN1 fibroblasts more vulnerable to cell death induced by oxidative stress. In this regard, G206D-PSEN1 mutation probably produces an autophagosome accumulation due to degradation defect. It induces a buildup and altered distribution of lysosomes, along with an increase of global p62/SQSTM1 that could compromise lysosomal degradation, as shown in Figure 7. All these alterations could contribute eventually to the neuronal degeneration that underlies the pathogenesis of Alzheimer’s disease. However, a limitation of our study is that the PSEN1 study is based on a single fibroblast cell line, and therefore, the conclusion drawn cannot be generalized; nevertheless, it opens the possibility of having mutation-specific treatments in the future.

This work constitutes an interesting characterization of the mitochondrial status and autophagy mechanisms in patients’ fibroblasts that could offer new targets for developing AD biomarkers and therapies.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by the Human Research Ethics Committees of CSIC and CIBERNED (Instituto de Salud Carlos III). The patients/participants provided their written informed consent to participate in this study.

PG-S and RM were responsible for all aspects of the project, including the conceptualization and design of the study. CV provided fibroblasts of AD patients. IC-L, PG-S, and EJ-E performed experiments and acquired and analyzed data. IC-L and PG-S provided the first draft, prepared figures, and discussed the results. PG-S, IC-L, EJ-E, CV, and RM edited the draft. PG-S provided the definitive version. RM and CV provided financial support. All authors revised the final version of the manuscript and took responsibility for its content.