Retinoic Acid Enhances Apolipoprotein E Synthesis in Human Macrophages

INTRODUCTION

Alzheimer’s disease (AD) is one of the most severe and least treatable neuropsychiatric disorders. The global burden of disease attributable to AD is still growing due to the worldwide demographical changes and the overall aging society [1].

Besides the well-known but rare monogenetic causes of AD, most patients (>90%) suffer from sporadic AD. In sporadic AD, the lipoprotein-coding Apolipoprotein E4 (APOE4) allele and the monocyte-receptor coding TREM2 are the most relevant known genetic risk factors [2]. While the overall lifetime risk for developing AD is about 12.5%, it increases to more than 50% for homozygous APOE4 carriers and ranges around 20–30% for APOE3/APOE4 heterozygotes [3].

The APOE gene encodes apolipoprotein E (ApoE) that is essentially involved in the homeostasis of certain lipids, mediating cholesterol flux to neurons [4]. Moreover, ApoE plays a significant role in amyloid-β (Aβ) homeostasis, regulating processes like neurite outgrowth and neuronal plasticity [5, 6], and critically mediates Aβ aggregation, clearance, and neurotoxicity [7–13]. Most interestingly, ApoE was shown to mediate the neuroprotective properties of microglia against Aβ-induced neurotoxicity in animal models [8], suggesting ApoE as a highly promising, potentially disease-modifying target in AD treatment and prevention.

While cerebral ApoE has long been thought to be solely a product of astrocytes, it is now generally accepted that microglia also synthesize relevant amounts of ApoE, thereby contributing to local ApoE homeostasis [8, 14–17]. In fact, a recent study even revealed microglia to synthesize up to 16 times more ApoE than astrocytes [14].

Increased microglial activation [18] is a hallmark in AD pathogenesis [19, 20] and has been shown to decrease glial ApoE synthesis and release [14–16]. This might result in a vicious cycle between neurotoxic inflammatory processes and decreased neuroprotective ApoE levels with microglial cells playing a central role.

Retinoic acid (RA), the active and highly neuroprotective metabolite of vitamin A, exhibits anti-inflammatory properties and reduces microglial activation [21, 22]. Furthermore, RA decreases the expression of β-secretase 1 (BACE1), a key enzyme in AD-related amyloid-β protein precursor (AβPP) processing, and attenuates the synthesis of the inflammatory cytokines IL-6 and TNFα [23]. This suggests a key role of RA in Aβ homeostasis and in the regulation of brain inflammatory processes. Further evidence supporting a central role of retinoids in AD pathology comes from findings on the mechanisms of action of minocycline, an anti-inflammatory and microglial-inhibitory substance that attenuates AD-related pathology in murine models [24–27]. Minocycline was shown to enhance RA-levels by inhibiting RA-degradation [22]. Moreover, defective Retinoic Acid Receptors (RAR) or Retinoic X Receptors (RXR) resulted in memory impairment in mice [28], while agonists at the RAR/RXR binding partners liver X receptor (LXR) and peroxisome proliferator-activated receptor γ (PPARγ) were shown to reduce Aβ levels and to improve cognitive function [29, 30]. Furthermore, the RXR-agonist bexarotene strikingly reduces Aβ burden, likely through upregulation of ApoE expression [9]. Most interestingly, anti-inflammatory and anti-apoptotic properties of bexarotene after traumatic brain injury were recently shown to be mediated through ApoE-related mechanisms [31].

While numerous studies have provided ample evidence for retinoid-mediated microglial ApoE homeostasis in AD pathology, there have been no studies on this likely important mechanism in humans to date. Therefore, we sought to assess systematically the impact of retinoid signaling and proinflammatory activation on ApoE homeostasis in primary, human-derived microglia-like macrophage cultures. Such a primary cell-based approach may be highly useful in the development of functional diagnostic and treatment-indicative biomarkers. The use of PBMCs is particularly useful, because this fraction of leukocytes contains the precursors of tissue-specific macrophages, the monocytes. These can easily be enriched and differentiated into macrophages, resulting in cells that phenomenologically closely resemble microglial cells [32]. In fact, it is well accepted that monocyte-derived microglial cells that can be found in the CNS of patients with neuropsychiatric disorders make up a significant proportion of all macrophages [32–34]. While most microglial cells appear to be yolk-sac derived, the monocyte-derived microglial-like cells do not differ functionally from those genuine microglial cells [35–37]. Thus, monocyte-derived macrophages differentiated in the presence of microglial-specific cytokines and growth factors, are used in our study to model patient-specific, microglial-like cells. Our investigations on retinoid-mediated ApoE synthesis in a personalized cell culture model provide a very first step toward the targeted development of functional biomarkers that mimic one of the most central pathogenetic mechanisms of this devastating disease.

METHODS

RESULTS

In the first step, we assessed whether inflammatory stimulation may, in analogy to findings in murine cell cultures, result in an attenuation of ApoE synthesis in PBMC-derived macrophages. While vehicle-treated cells exhibited baseline ApoE levels of 4.4±1.9 ng/ml, inflammatory stimulation with LPS (10 ng/ml) for 24 h resulted in reduced intracellular ApoE levels (2.2±0.9 ng/ml), which were assessed after harvesting of the cells 24 h upon addition of LPS. However, this effect did not reach statistical significance (p = 0.65) (Fig. 1).

OPEN IN VIEWER

Fig.1

Effect of inflammatory stimulation and retinoic acid (RA) exposure on ApoE levels in human macrophages. ApoE is expressed in human PBMC-derived macrophages. Inflammatory stimulation with lipopolysaccharide (LPS; 10 ng/mL) reduces ApoE levels in human PBMC-derived macrophages. Treatment with RA (10 μM) led to an increase in intracellular ApoE that was not affected by additional LPS-treatment. ApoE levels were measured 24 h after treatment. Results are means ± SEM from 3 healthy controls assessed in 3 independent experiments. ^∗∗significantly different to VEH, p<0.001.

Next, we studied the moderating role of RA signaling in microglial ApoE synthesis. Microglial-like human macrophage cell cultures were treated with RA (10 μM), resulting in a strong, significant increase of intracellular ApoE levels (22.02±2.9, p < 0.001).

To assess the effect of RA on activated microglial-like macrophages, cells were treated with RA and with pro-inflammatory LPS (or medium for sham condition) for 24 h. RA completely abolished the effect of LPS on intracellular ApoE levels (24.7±3, p < 0.001) (Fig. 1).

DISCUSSION

Here, we show for the first time that human PBMC-derived macrophages, which are functionally related to genuine adult microglia, are capable of synthesizing ApoE and respond to proinflammatory activation in a similar pattern as previously published for murine microglia [14]. These findings may represent a first step toward a novel and unique personalized model to measure patient-specific microglial ApoE homeostasis in response to, e.g., inflammatory stimuli and other endogenous or exogenous compounds. Since chronic, sub-threshold neuroinflammatory processes are discussed to underlie AD pathology, our findings on LPS-disrupted ApoE synthesis and the abolishment of this effect by RA, provide first evidence for a central, possibly disease-modifying pathway.

Microglia are the primary immune cells of the CNS and functionally highly related to peripheral macrophages [36]. Microglia in the brain are even partially recruited from circulating blood monocytes [32–34, 38] and express numerous macrophage receptors [35]. Thus, patient specific macrophages that are differentiated in the presence of factors required for microglial differentiation may be suited as an easily accessible cell model for genuine, brain-derived microglia. Usually, astrocytes alone are considered to be the major source for ApoE in the CNS. However, several animal models have shown a significant contribution of microglia to local ApoE synthesis. Saura et al. demonstrated that murine microglial cell cultures exhibited an up to 2-fold higher ApoE synthesis rate than mixed glial cultures [15]. Polazzi et al. even found an up to 16-fold higher ApoE expression in microglia compared to astrocytes [14].

Reduced ApoE synthesis upon LPS exposure has repeatedly been demonstrated in murine microglia [15, 16], but to our knowledge has never been shown in human microglial (-like) cells before. Although our data on LPS-attenuated ApoE synthesis did not reach statistical significance, the overall effect of rather small amounts of LPS (10 ng/ml) suggests the presence of inflammation-induced ApoE reduction in human cell cultures. Since we only included cells derived from three healthy subjects, interindividual variability may be the cause for the heterogeneous results, suggesting the amount of inflammation-induced ApoE attenuation as a promising functional biomarker.

A pathological pro-inflammatory activation of microglia is known for subjects with AD [19], whereas a blockage of microglial activation was shown to prevent memory loss in mice [39]. The importance of microglia in the pathogenesis of AD is partly due to their ability to participate in the elimination (phagocytosis) of neurotoxic Aβ peptides, one of the main pathological factors in AD. Here, we reveal another pathway between neuro-inflammation and microglia-associated AD pathology, as our results demonstrate that proinflammatory stimulation with LPS decreases microglial ApoE synthesis. Especially against the background that ApoE is known to influence microglial Aβ uptake [40, 41], our results point towards this vicious cycle between inflammation, ApoE synthesis and Aβ elimination and neurodegeneration with microglia in the central role (Fig. 2).

OPEN IN VIEWER

Fig.2

Vicious cycle between inflammation, ApoE synthesis, and Aβ elimination and neurodegeneration with microglia in the central role. Microglial activation leads to a decreased microglial ApoE synthesis and a reduction of microglial Aβ uptake. These results in neurodegeneration in two ways: decreased neuroprotection and increased neurotoxicity. RA, that inhibits microglial activation, might have the potential to interrupt this vicious cycle between neuro-inflammation, microglia activation, ApoE secretion and Aβ homeostasis.

Highlighting the central role for RA signaling in ApoE homeostasis, Zhao et al. have previously demonstrated a 4-fold increase of ApoE secretion in murine astrocytes upon RA exposure [42]. Our results demonstrate a similar increase in macrophageal ApoE levels upon RA treatment.

With respect to the underlying mechanisms, we have previously demonstrated a role for RA in inhibiting LPS-induced microglial activation, which itself is a trigger for increased RA catabolism [22], giving one possible explanation for the effects of LPS on ApoE synthesis, and its reversal through RA exposure.

Taken together, RA may act as a key linkage between AD-related chronic, sub-threshold neuro-inflammation, associated microglial activation, eventually increased RA turnover and subsequently attenuated ApoE secretion. Finally, reduced ApoE levels may impair Aβ-phagocytosis [9], resulting in the canonical plaque deposition and eventually tangle formation with subsequent neurodegeneration.

In summary, our data show for the first time that RA strikingly increases ApoE synthesis in human primary blood-derived cell cultures. Inflammation-induced attenuation of ApoE synthesis can be rescued through RA treatment. Our findings provide first evidence for retinoids as a highly promising, possibly disease-modifying strategy. With respect to our personalized cell culture model we will be able to model these aspects at a patient-specific level.

DISCLOSURE STATEMENT

Authors’ disclosures available online (https://www.j-alz.com/manuscript-disclosures/17-0823r1).