Female sex is linked to a stronger association between sTREM2 and CSF p-tau in Alzheimer’s disease

Beta-amyloid (Aβ) and tau are the hallmark pathologies of Alzheimer’s disease (AD), ensuing neurodegeneration and cognitive decline (Jack et al, 2018). Previously it was shown that Aβ-oligomers trigger the neuronal secretion of soluble phosphorylated tau (p-tau) (Jin et al, 2011), which in turn drives the aggregation and trans-synaptic spread of neurofibrillary tau tangle pathology (Pichet Binette et al, 2022). These findings indicate that Aβ-related p-tau increases are critical for the subsequent development of tau aggregates. While the sequence of the amyloid cascade is well-established (Glenner and Wong, 1984; Hardy and Selkoe, 2002; Selkoe and Hardy, 2016), the progression rate of Aβ-initiated tau accumulation differs significantly between patients, causing heterogenous disease trajectories and dynamics across patients (Dujardin et al, 2020; Komarova and Thalhauser, 2011; Landau et al, 2022). The aim of the present study is therefore to determine modulating factors of the Aβ to p-tau axis, which can help identify factors that accelerate or attenuate tau progression and determine targets to attenuate p-tau secretion and subsequent tau aggregation. To this end, we specifically focused on factors that have been previously associated with an increased risk of tau pathology in AD, which are female sex, microglial activation, younger age, and ApoE4, which is the main genetic risk factor for AD.

Accumulating evidence indicates that women are more severely affected by AD than men (Buckley et al, 2019b; Fisher et al, 2018; Laws et al, 2018; Levine et al, 2021; Mosconi et al, 2017; Nebel et al, 2018; Vest and Pike, 2013), accounting for two-thirds of AD dementia cases in the US (Alzheimer’s, 2014). Previous work revealed that women show increased levels of CSF total tau and p-tau (Hohman et al, 2018), tau deposition (Buckley et al, 2019b), and a faster tau accumulation rate compared to men (Smith et al, 2020). Recently it was found that faster tau accumulation in women was facilitated by a stronger association between Aβ fibrils and soluble p-tau in women compared to men, suggesting early Aβ-dependent tau secretion as a critical turnover point for the observed sex differences in AD (Wang et al, 2024). Although the driving mechanisms behind these findings are not fully clear, differences in sex hormones (Sundermann et al, 2020) and inflammatory processes (Casaletto et al, 2022) are assumed to play an underlying role in the manifestation of sex differences in AD. In this context, analysis of post-mortem brain tissue of older adults show that microglial activation mediated the association between Aβ plaque pathology and neurofibrillary tau tangles in women but not in men (Casaletto et al, 2022). An association between microglial activation and the accumulation and spread of tau in AD has been previously shown in vitro (Brelstaff et al, 2021; Maphis et al, 2015) and in vivo (Pascoal et al, 2021; Vogels et al, 2019), thus microglial activation may play an early key role in the amyloid cascade (Pascoal et al, 2021). From a mechanistic point of view, microglia are the brains innate immune cells, which react to hazardous stimuli with the release of pro-inflammatory cytokines. Once the hazard stimulus has been eliminated, microglia return to their original homeostatic resting state. However, in chronic inflammation, microglia lose their capability to return back to their homeostatic state, resulting in neurotoxicity and tissue damage (Bivona et al, 2023). The Triggering Receptor Expressed on Myeloid Cell 2 (TREM2) has been associated with a shift of microglia from homeostatic to disease associated states (Keren-Shaul et al, 2017; Krasemann et al, 2017), parallels PET-assessed microglial activation (Brendel et al, 2017), and is thus a well-established proxy for microglial activation (Ewers et al, 2020; Ewers et al, 2019; Franzmeier et al, 2020). We and others showed recently that increased CSF sTREM2 levels are associated with higher levels of p-tau in early phases of sporadic AD (Suarez-Calvet et al, 2019), and mediate Aβ-related p-tau increases in earliest Aβ fibrillization (Biel et al, 2023). Thus, microglial activation might have a pivotal role in the early pathogenesis of AD, with different effects in men and women.

A further potential modulator of the Aβ to p-tau axis might relate to the patient’s age. In addition to the effects of sex and sTREM2-related microglial activation, studies show that younger age at symptom onset is associated with a worsened prognosis and faster tau accumulation in sporadic AD (Frontzkowski et al, 2022; Koedam et al, 2008; van der Vlies et al, 2009). However, it remains unclear, whether younger age modulates the Aβ to p-tau axis towards a faster p-tau increase, which would support the view of a more aggressive form of AD when patients enter the amyloid cascade at younger age (Koedam et al, 2008; Touroutoglou et al, 2023).

Finally, genetic predispositions might play a critical role for early tau pathology. The ε4 allele of the ApoE-encoding apolipoprotein (ApoE4) is the main genetic risk factor for sporadic AD, and carriership of the ApoE4 risk allele has been linked to an oversupply of cholesterol, ensuing accelerated Aβ production (Lee et al, 2021), neuroinflammation (Ophir et al, 2005), impaired myelination (Blanchard et al, 2022), and tau-mediated neurodegeneration (Shi et al, 2017). Further, ApoE4 carriership is associated with lower levels of testosterone (Hogervorst et al, 2002), a sex hormone which is mostly expressed in males and associated with anti-inflammatory processes (Bianchi, 2019; Ota et al, 2012) and cholesterol clearance (Kilby et al, 2021). Previously, we showed that ApoE4 drives Aβ-related tau accumulation at lower levels of Aβ pathology, suggesting that in ApoE4 carriers, tau accumulation starts earlier than in ApoE4 non-carriers (Steward et al. 2023). In addition, several studies indicate that female ApoE4 carriers show enhanced levels of soluble CSF total tau (Altmann et al, 2014; Babapour Mofrad et al, 2020; Buckley et al, 2019a; Damoiseaux et al, 2012; Hohman et al, 2018) and p-tau (Babapour Mofrad et al, 2020; Hohman et al, 2018) compared to their male counterparts, hence, the association between ApoE4 carriership and Aβ-dependent tau might be further modulated by sex.

Together, female sex, microglial activation, younger age, and ApoE4-related genetic predisposition for AD have been previously associated with increased tau burden. The main aims of the present study were therefore to test, (i) whether sex, sTREM2-related microglial activation, younger age or ApoE4 modulate the Aβ to p-tau axis, and (ii) whether the effects of sTREM2, younger age or ApoE4 are stronger in women than in men.

In the present study, we systematically assessed modulating factors of the Aβ to p-tau axis, i.e., a potentially critical driver of tau fibrillization in AD (Pichet Binette et al, 2022), to better understand heterogeneity in pathophysiological disease progression. Specifically, we assessed whether sex, microglial activation (i.e., sTREM2), age or genetic predisposition for AD (i.e., ApoE4) are associated with higher levels in p-tau₁₈₁, and specifically, modulate the Aβ to p-tau axis. Since women are at increased risk of AD, we further tested whether effects of any potential Aβ to p-tau modulator are stronger in women than in men. First, we show that sTREM2 and ApoE4 are associated with higher p-tau₁₈₁ levels and that the effects were even more pronounced in women than in men. Second, we show that solely sex modulates Aβ-dependent p-tau₁₈₁ levels, with stronger Aβ-dependent p-tau₁₈₁ secretion in women compared to men. Finally, we observed a trend for sex differences for the association between sTREM2 and Aβ-dependent p-tau₁₈₁ levels, again, with higher p-tau₁₈₁ levels in women than in men. Together, our results underline sex-specific dynamics in AD disease pathways with more severe consequences for higher p-tau₁₈₁ levels in women. The findings are critical for patient stratifications in clinical trials, especially for drugs targeting microglial activation as a disease modifying approach.

Previously, higher levels of soluble (Hohman et al, 2018; Tsiknia et al, 2022) and aggregated tau (Buckley et al, 2019b; Shokouhi et al, 2020) were found in women compared to men. Congruently, we found that women showed stronger Aβ-related p-tau₁₈₁ secretion and trend to have faster Aβ-related p-tau₁₈₁ increases than men (p_FDR = 0.092), with sex being the only tested factor that modulated the Aβ to p-tau axis. However, future studies are needed to further investigate the effect of sex on Aβ-related longitudinal p-tau propagation. Besides a main effect of sTREM2-related microglial activation on p-tau₁₈₁ levels, we further revealed a stronger association between sTREM2 and p-tau₁₈₁ in women than in men. When testing whether this observation could be applied to the Aβ to p-tau axis, we found a 3-way interaction of Aβ, sTREM2, and sex on p-tau₁₈₁, showing that in women, higher levels of sTREM2 were linked to a stronger Aβ-dependent p-tau₁₈₁ response. However, when applying FDR correction, the result only reached borderline significance (p_FDR = 0.063), hence, future studies are needed to confirm this finding. In our previous work investigating disease stage-dependent effects of sTREM2-related microglial activation on p-tau₁₈₁ increases, we found that in patients within earliest Aβ pathology (defined as Aβ CSF positive and amyloid-PET negative) (Palmqvist et al, 2017), sTREM2 mediated Aβ-related p-tau₁₈₁ increases (Biel et al, 2023). In addition, early Aβ pathology was associated with glucose hypermetabolism, indicating that sTREM2 follows earliest Aβ fibrillization, which might manifest in activated microglia consuming more glucose (Xiang et al, 2021). Our current observations align well with findings using TSPO-PET as a proxy of microglial activation, where TSPO-PET was associated with higher tau-PET signals in female but not in male AD patients (Biechele et al, 2024). Similarly, in post-mortem investigations, microglial activation was linked to Aβ-related tau pathology in women but not in men (Casaletto et al, 2022). Hence, microglial-induced tau accumulation might be more pronounced in women than in men. Biologically, women are predisposed to higher levels of neuroinflammatory markers than men, which has been shown using TSPO-PET across studies in healthy adults (Tuisku et al, 2019) and animal models of amyloidosis (Biechele et al, 2020). Preclinical research found that microglia have transcriptional sex differences in adult brains, suspected to be caused by sex chromosomes as well as sex hormones that might be involved in microglial functioning (Guillot-Sestier et al, 2021; Kodama and Gan, 2019; Villa et al, 2018). In addition, women are more often affected by autoimmune diseases of the central nervous system, such as multiple sclerosis (Kalincik et al, 2013; Koch-Henriksen and Sorensen, 2010), which has been further attributed to maladaptive microglial activation (Yong, 2022). Thus, sex-specific differences in microglial activation in women might result in a higher vulnerability to chronic neuroinflammation, which may cause neuronal damage (Bivona et al, 2023; Jayaraman et al, 2021). It would be critical to assess, whether the observed association between sTREM2-related microglial activation and p-tau₁₈₁ is mediated by pro-inflammatory cytokines (e.g., Interleukin 1β), which might result from prolonged microglial activation as a response to earliest Aβ fibrillization (Wang et al, 2015). Here, it should be tested, whether the threshold for a pro-inflammatory cytokine response differs between men and women. In addition, future studies should include additional markers related to neuroinflammation in the CSF (e.g., YKL-40, ICAM-1, VCAM1) (Janelidze et al, 2018; Popp et al, 2017) or tissue (e.g., TSPO-PET, FDG-PET) (Xiang et al, 2021) to address the underlying mechanism that link neuroinflammation and p-tau secretion. Moreover, future investigations could benefit from incorporating proteomic analyses to further explore more complex patterns of neuroinflammation markers, particularly those related to microglial activation. Detrimental effects of microglial activation in early disease stages of AD are in contrast to the observed positive effects microglial activation might have in later stages of AD, such as protective effects on Aβ pathology, neurodegeneration, and cognitive decline (Ewers et al, 2020; Ewers et al, 2019; Morenas-Rodriguez et al, 2022). We therefore suggest to consider subanalyses in clinical trials that are stratified by sex and disease stage, as our findings indicate that men and women show different dynamics in AD disease pathways. These might be driven among others by a different response to microglial activation and presumably earlier thresholds for Aβ-related neuroinflammation in women than in men.

Besides the association between sTREM2-related microglial activation and p-tau₁₈₁, we observed a main effect of ApoE4 status on cross-sectional p-tau₁₈₁ as well as longitudinal p-tau₁₈₁ increases, which supports our previous work showing that ApoE4 enhances tau spreading using tau-PET (Steward et al, 2023). From a mechanistic point of view, it has been shown that the ApoE4 allele interferes with the absorption of polyunsaturated fatty acids, which are vital for the cell’s functioning. As a consequence of ApoE4, fewer nutrients can be absorbed, the cells become inflamed and ultimately die (Asaro et al, 2020). This inflammatory reaction might promote the secretion of soluble tau, similar to the observed effects of sTREM2-related microglial activation on p-tau (Biel et al, 2023). Indeed, previous work reported that ApoE4 facilitates microglia-related neuroinflammation and thereby might contribute to AD pathways (Kang et al, 2018; Krasemann et al, 2017; Parhizkar and Holtzman, 2022; Tai et al, 2015; Ulrich et al, 2018). Specifically, it was recently shown that ApoE4 activates microglia within brain regions that are prone to early tau propagation, and this effect was independent of Aβ (Ferrari-Souza et al, 2023). In addition, and in line with previous observations (Altmann et al, 2014; Babapour Mofrad et al, 2020; Damoiseaux et al, 2012; Hohman et al, 2018), we found that female ApoE4 risk allele carriers show higher levels of cross-sectional p-tau₁₈₁ compared to male ApoE4 risk allele carriers. Importantly, higher microglia-induced inflammatory states were previously found in female ApoE4 carriers compared to male ApoE4 carriers (Mhatre-Winters et al, 2022), suggesting similar sex-specific associations between sTREM2- and ApoE4-related neuroinflammation and p-tau₁₈₁ levels. However, with the data of the present study, causal conclusions are limited, thus, future work is needed to test the link between ApoE4-induced cell inflammation and subsequent p-tau secretion. Further, our finding of increased p-tau₁₈₁ in ApoE4 carriers may reflect, in part, their predisposition toward earlier Aβ pathology onset and thus a more advanced disease stage (Therriault et al, 2021). Given that ApoE4 carriers in our study show higher Aβ positivity and load (Figs. 1 and 2), the observed association with p-tau₁₈₁ could be influenced by their progression along the amyloid cascade. Future longitudinal studies with stage-specific controls would help clarify whether this relationship is independent of ApoE4’s effects on Aβ progression. In contrast, ApoE4 status did not modulate the association between Aβ and p-tau₁₈₁, which seems surprising since ApoE4 has been extensively identified as a driver of Aβ pathology (Liu et al, 2017; Morris et al, 2010; Reiman et al, 2009). Similarly, the interaction between ApoE4 and sex on p-tau₁₈₁ was no longer present when including Aβ as interaction term. Recently, we found that Aβ mediates the association between ApoE4 and faster tau accumulation in regions that are vulnerable for early tau aggregation (Steward et al, 2023), thus, ApoE4 might only drive Aβ-related p-tau increases in early AD disease stages, while in later disease, the effects of ApoE4 on p-tau might be independent of Aβ fibrillization. Along the same lines, it has been noted that the interaction between ApoE4 and sex on p-tau levels is only persistent in early disease stages (subjective cognitive decline and MCI, but not in dementia) (Babapour Mofrad et al, 2020), consistent with another study reporting an ApoE4 x sex interaction on tau-PET only within early regions of tau deposition (Wang et al, 2021). In the context of microglial activation and neuroinflammation, we also previously showed that sTREM2 mediates Aβ-related p-tau₁₈₁ increases only in early Aβ fibrillization (Biel et al, 2023). It would be a key next step to test whether the observed sex interactions with sTREM2 and ApoE4 on p-tau₁₈₁ are disease stage-dependent and whether they are related to each other.

Finally, we did not observe any effects of age on cross-sectional or longitudinal p-tau₁₈₁ levels, neither a main effect nor an interaction with Aβ or sex. Older age is the major risk factor for developing AD (Guerreiro and Bras, 2015; Hou et al, 2019). However, younger age at symptom onset has been conversely found to be associated with accelerated tau accumulation (Smith et al, 2020), neurodegeneration (Moller et al, 2013), cognitive decline (van der Vlies et al, 2009) and higher rates of mortality in AD (Koedam et al, 2008). Therefore, we tested whether accelerated spread of tau in younger patients can be further explained by a stronger association between Aβ and p-tau₁₈₁, which was not confirmed by our analyses. Therefore, faster tau accumulation in younger AD patients is unlikely driven by a higher Aβ-related p-tau response and it is unclear why tau accumulates in younger patients at a faster rate and whether female patients with an early onset show a faster spreading of tau. Here, future research should further address the underlying mechanisms of tau accumulation in patients with an earlier disease onset.

A strength of this study is the inclusion of several biomarker assessments in relation to sex differences in AD, investigating modulating effects on the Aβ to p-tau axis within a large sample ranging from cognitively normal to demented. However, several limitations should be addressed when interpreting our data. First, sTREM2 data have been obtained relatively late in the ADNI dataset, hence, data with subsequent CSF p-tau₁₈₁ for assessing longitudinal relationships were limited. Since we previously show that sTREM2 mediated the association between Aβ and p-tau₁₈₁ only in earliest Aβ fibrillization, we restricted the analysis on cross-sectional data for the sTREM2 sample in order to increase the sample covering Aβ− controls to Aβ+ patients across the AD spectrum to 454 participants. We encourage future studies to assess sex differences in longitudinal associations between Aβ, sTREM2, and p-tau₁₈₁ once more data are available. Second, sTREM2 is only an indirect marker of microglial activation, thus other direct (e.g., post-mortem) or indirect (e.g., TSPO-PET) marker should be included in future studies. Third, in the present study, not sufficient tau-PET data were available to reliably test whether the stronger p-tau₁₈₁ response in women promotes faster spreading of tau aggregates. Once more data are available, this might be an important target for future investigations. Finally, the analysis should be replicated in other cohorts than ADNI and include more diverse participant groups with different ethnicities in order to increase generalizability of our findings.

In conclusion, our findings show that sex is an important modulator of Aβ-dependent p-tau₁₈₁ secretion. In particular, women show a stronger association between sTREM2-related microglial activation and p-tau₁₈₁ than men, supporting the view that neuroinflammation may play a key role in the observed sex differences in AD. In addition, we found higher p-tau₁₈₁ levels in female ApoE4 carriers compared to male ApoE4 carriers, which might be as well attributable to greater neuroinflammatory effects of ApoE4 in women. Our study provides evidence for sex-specific dynamics of AD disease pathways related to neuroinflammation and has potential implications for drug treatments targeting microglial activation, in which sex should be considered as an important modulating factor.

All data used in this manuscript are publicly available from the ADNI database (adni.loni.usc.edu) upon registration and compliance with the data use agreement. The data that support the findings of this study are available upon reasonable request from the corresponding author.

The source data of this paper are collected in the following database record: biostudies:S-SCDT-10_1038-S44321-024-00190-3.

Davina Biel: Conceptualization; Data curation; Formal analysis; Investigation; Writing—original draft; Writing—review and editing. Marc Suárez-Calvet: Writing—review and editing. Anna Dewenter: Writing—review and editing. Anna Steward: Writing—review and editing. Sebastian N Roemer: Writing—review and editing. Amir Dehsarvi: Writing—review and editing. Zeyu Zhu: Writing—review and editing. Julia Pescoller: Writing—review and editing. Lukas Frontzkowski: Writing—review and editing. Annika Kreuzer: Writing—review and editing. Christian Haass: Data curation; Writing—review and editing. Michael Schöll: Writing—review and editing. Matthias Brendel: Conceptualization; Writing—review and editing. Nicolai Franzmeier: Conceptualization; Data curation; Formal analysis; Investigation; Writing—original draft; Writing—review and editing.

Source data underlying figure panels in this paper may have individual authorship assigned. Where available, figure panel/source data authorship is listed in the following database record: biostudies:S-SCDT-10_1038-S44321-024-00190-3.

Open Access funding enabled and organized by Projekt DEAL.

The authors declare no competing interests.

Data used in preparation of this article were obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). As such, the investigators within the ADNI contributed to the design and implementation of ADNI and/or provided data but did not participate in analysis or writing of this report. A complete listing of ADNI investigators can be found at: http://adni.loni.usc.edu/wp-content/uploads/howtoapply/ADNIAcknowledgementList.pdf. We thank the ADNI participants and their families who made this study possible. The study was funded by grants from the LMU (FöFoLe, 1119, awarded to DB), the Hertie foundation for clinical neurosciences (awarded to NF), and the SyNergy excellence cluster (EXC 2145/ID 390857198).