Beneficial effects of increased lysozyme levels in Alzheimer's disease modelled in Drosophila melanogaster

Genetic polymorphisms of immune genes that associate with higher risk to develop Alzheimer's disease (AD) have led to an increased research interest on the involvement of the immune system in AD pathogenesis. A link between amyloid pathology and immune gene expression was suggested in a genome‐wide gene expression study of transgenic amyloid mouse models. In this study, the gene expression of lysozyme, a major player in the innate immune system, was found to be increased in a comparable pattern as the amyloid pathology developed in transgenic mouse models of AD. A similar pattern was seen at protein levels of lysozyme in human AD brain and CSF, but this lysozyme pattern was not seen in a tau transgenic mouse model. Lysozyme was demonstrated to be beneficial for different Drosophila melanogaster models of AD. In flies that expressed Aβ1‐42 or AβPP together with BACE1 in the eyes, the rough eye phenotype indicative of toxicity was completely rescued by coexpression of lysozyme. In Drosophila flies bearing the Aβ1‐42 variant with the Arctic gene mutation, lysozyme increased the fly survival and decreased locomotor dysfunction dose dependently. An interaction between lysozyme and Aβ1‐42 in the Drosophila eye was discovered. We propose that the increased levels of lysozyme, seen in mouse models of AD and in human AD cases, were triggered by Aβ1‐42 and caused a beneficial effect by binding of lysozyme to toxic species of Aβ1‐42, which prevented these from exerting their toxic effects. These results emphasize the possibility of lysozyme as biomarker and therapeutic target for AD.

Genetic polymorphisms of immune genes that associate with higher risk to develop Alzheimer's disease (AD) have led to an increased research interest on the involvement of the immune system in AD pathogenesis. A link between amyloid pathology and immune gene expression was suggested in a genome-wide gene expression study of transgenic amyloid mouse models. In this study, the gene expression of lysozyme, a major player in the innate immune system, was found to be increased in a comparable pattern as the amyloid pathology developed in transgenic mouse models of AD. A similar pattern was seen at protein levels of lysozyme in human AD brain and CSF, but this lysozyme pattern was not seen in a tau transgenic mouse model. Lysozyme was demonstrated to be beneficial for different Drosophila melanogaster models of AD. In flies that expressed Ab  or AbPP together with BACE1 in the eyes, the rough eye phenotype indicative of toxicity was completely rescued by coexpression of lysozyme. In Drosophila flies bearing the Ab 1-42 variant with the Arctic gene mutation, lysozyme increased the fly survival and decreased locomotor dysfunction dose dependently. An interaction between lysozyme and Ab 1-42 in the Drosophila eye was discovered. We propose that the increased levels of lysozyme, seen in mouse models of AD and in human AD cases, were triggered by Ab  and caused a beneficial effect by binding of lysozyme to toxic species of Ab  , which prevented these from exerting their toxic effects. These results emphasize the possibility of lysozyme as biomarker and therapeutic target for AD.

Introduction
Two of the main pathological hallmarks of Alzheimer's disease (AD) are extracellular accumulation of amyloid plaques, which consists of amyloid-b (Ab) peptides, and intracellular neurofibrillary tangles composed of hyperphosphorylated tau [1]. The amyloid plaques were previously considered the foremost neurotoxic agents in AD, but increasing evidence suggests that small diffusible Ab aggregates (referred to as oligomers) are the principal cytotoxic species as these correlate better with synaptic loss and cognitive impairment than the plaques [2]. Ab peptides are produced from sequential cleavage of the amyloid-b precursor protein (AbPP) by b-secretase activity of the b-site AbPP-cleaving enzyme 1 (BACE1) and the csecreatase complex, where presenilin 1 (PSEN1) is the catalytic subunit [3]. Although Ab 1-40 is the predominant secreted form, the longer Ab 1-42 has a more prominent part in AD as it is more aggregation prone and forms toxic oligomers more easily [4]. Besides Ab and tau, it is well established that neuroinflammation is involved in AD. Activated astrocytes and microglia cells surround Ab plaques together with various inflammatory mediators [5,6]. In addition, genetic studies show upregulation of several genes involved in inflammation, especially complement activation and prostaglandin synthesis, during incipient AD [7] and there is data which demonstrate inflammatory processes before tangles and neurodegeneration are apparent [8]. However, microglial activation also demonstrates a protective function of a triggered immune response in AD, as microglial activation mediates phagocytosis and clearance of Ab [9].
Lysozyme, which belongs to the innate immune system, is upregulated in cerebrospinal fluid (CSF) from AD patients and inhibits the formation of toxic Ab oligomers [10][11][12]. Lysozyme overexpression in Ab transgenic Drosophila melanogaster rescues both the survival and the activity of the Ab flies [10]. Lysozyme is a glucoside hydrolase able to hydrolyse peptidoglycans found in the cell walls of bacteria [13]. It is secreted from macrophages and microglia, and it is abundant in various secretions such as tears, saliva, milk and CSF [14]. The aim of this study was to further investigate the implication of lysozyme in AD. Lysozyme gene expression was investigated using a database of a genome-wide gene expression study of wild-type (WT) and five mouse models of AD (mutant human AbPP, mutant human PSEN1, homozygous and heterozygous expressed AbPP-PSEN1 and mutant human TAU) [15], and a database of AD patient brain tissue [16]. The levels of lysozyme protein were investigated in brain tissue from transgenic AD mice and AD patients. An increased lysozyme expression was found both at mRNA and protein level in AD brain tissue of both mice and humans. In order to investigate the impact of lysozyme expression during AD, three different Drosophila models of AD were used. Beneficial effects of lysozyme in these different Drosophila models were discovered; in flies that expressed Ab 1-42 individually or AbPP together with BACE1 (AbPP-BACE1) in the fly eyes, the AD phenotype was completely rescued by lysozyme. In flies carrying the highly toxic Ab peptide with the Arctic mutation (E22G), lysozyme increased the fly survival and improved the locomotor behaviour in a dose-dependent manner. These results imply that lysozyme has a protective effect on Ab toxicity and could function as a new therapeutic strategy for AD.

Lysozyme is increased in brains of transgenic AD mice
To investigate whether the mRNA expression of lysozyme is changed during AD progression, we used data from the publicly available database www.mouseac. com on five different amyloid or tau mouse dementia models. The mouse models were analysed at the ages 2, 4, 8 and 18 months [15]. Homozygous and heterozygous expression of human AbPP, with the Swedish mutation in combination with mutant human PSEN1 (AbPP-PSEN1), leads to plaque formation at 4 and 8 months, respectively, mutant AbPP expressed separately leads to plaques first at 18 months and mutant PSEN1 expressed separately has no plaque pathology. The mutant human heterozygous TAU mice demonstrate tangles at 8 months. The gene expression of lysozyme in the homozygous AbPP-PSEN1 mice was found to be significantly increased at 4 months in cortex (Fig. 1A) and hippocampus (Fig. 1B) and in heterozygous AbPP-PSEN1 mice at 8 months compared to WT mice (Fig. 1A,B). Lysozyme levels were unchanged in cerebellum of both homozygous and heterozygous AbPP-PSEN1 mice (Fig. 1C). In AbPP mice, there was a trend of increased lysozyme gene expression in cortex at 18 months, but not in hippocampus and no change was detected in PSEN1 mice ( Fig. 1A-C). We next investigated the correlation between lysozyme gene expression and Ab pathology in the cortex (Fig. 1G) and hippocampus (Fig. 1H) of these mice. Both heterozygous and homozygous AbPP-PSEN1 mice showed a strong and significant linear correlation in the cortex (r = 0.91 and 0.94 respectively) and in  hippocampus (r = 0.86 and 0.95 respectively) (Fig. 1G,H). Mice only expressing AbPP exhibited a strong and significant correlation as well, both in cortex and hippocampus, albeit weaker than for the double transgenic mice (r = 0.74 and 0.77 respectively) ( Fig. 1G,H). Tau transgenic mice showed an increase in lysozyme expression in cortex (Fig. 1D) and hippocampus ( Fig. 1E) only after 18 months. In the cerebellum of the tau transgenic mice, no change in lysozyme gene expression was detected (Fig. 1F). A sigmoid correlation relationship between lysozyme levels and tau pathology was demonstrated in hippocampus ( Fig. 1I; R 2 = 0.81), while this sigmoid correlation relationship was weaker in cortex ( Fig. 1I; R 2 = 0.52). These results demonstrate that the increase of Ab plaque pathology can trigger lysozyme expression in the AD mice brain, while tau pathology has no immediate impact on lysozyme expression. Next, lysozyme protein levels in WT and transgenic mice, harbouring PS1 and the AbPP Swe mutations, were investigated using western blot. AbPP-mutated mice had an increased protein expression of Ab and of lysozyme as compared to WT mice, which is consistent with the mRNA expression analysis of lysozyme ( Fig. 2A,B). To study whether lysozyme and Ab colocalized in plaques, immunohistochemistry was performed on frozen tissue from mice with the AbPP Swe mutation. Double-labelling with antibodies against lysozyme and Ab showed that lysozyme localized within the plaque as scattered dots (Fig. 2C).

Lysozyme is increased in the human AD brain
To study if the increased lysozyme mRNA expression levels in AD mice also applied to human AD patients, a database with data from autopsied tissues from visual cortex, dorsolateral prefrontal cortex and cerebellum was used [16]. The mRNA expression of lysozyme was significantly increased in visual cortex and prefrontal cortex of AD patients compared to control, but no difference was seen in cerebellum (Fig. 3A). To study if the protein expression levels of lysozyme were affected in AD, human post-mortem tissue from temporal cortex was used. The level of lysozyme was significantly increased in the AD group (Braak stages V-VI) compared to controls (Braak stages 0-IV) (Fig. 3B), which corresponds with the lysozyme mRNA expression analysis. In addition, lysozyme levels were investigated in an AD validated cohort of CSF and the lysozyme levels were significantly higher in CSF samples from AD patients compared to controls (Fig. 3C).

Lysozyme protects from Ab-induced toxicity in transgenic AD flies
In order to study if lysozyme has any effect on Ab toxicity, we used one Drosophila model where human lysozyme was coexpressed with Ab 1-42 and one, recently characterized, novel Drosophila model that coexpressed AbPP and BACE1 (AbPP-BACE1) [17] with or without lysozyme in the eyes of the flies using the retina-specific gmr-Gal4 driver. The eye morphology was examined at the day of eclosion using SEM. Ab 1-42 flies had a disturbed pattern of ommatidia compared to control flies (only expressing Gal4) and lysozyme flies (Fig. 4A,B). This disturbed ommatidia phenotype was completely rescued when Ab 1-42 was coexpressed with lysozyme ( Fig. 4A,B). AbPP-BACE1 flies showed a heavily disturbed pattern of ommatidia with fused and irregular-shaped ommatidia (Fig. 4A, B), while flies that expressed AbPP demonstrated a regular and symmetric pattern of ommatidia and BACE1 flies exhibited a small disruption of the eye phenotype (Fig. 4A,B). Coexpression of lysozyme with the AbPP-BACE1 flies significantly rescued the disturbed ommatidia phenotype (Fig. 4A,B).

Lysozyme does not change the level of Ab but binds to Ab in the Drosophila eye
Having established that lysozyme has a protective effect in the AD transgenic flies, we next investigated if coexpression of lysozyme in these flies could change the level of soluble and insoluble Ab 1-42 . The level of Ab 1-42 in the heads of the flies was measured by the Meso Scale Discovery technique. There was a significant increase in both soluble and insoluble levels of Ab 1-42 in flies that expressed Ab  or AbPP and BACE1 compared to their controls (only expressing Gal4 or AbPP respectively) (Fig. 4D,E). The proportion of insoluble and soluble Ab 1-42 were compared in the flies and the level of insoluble Ab 1-42 was eight times higher than soluble Ab 1-42 in the Ab 1-42 flies and the level of insoluble Ab 1-42 was 40 times higher than soluble Ab 1-42 in the AbPP-BACE1-expressing flies. Coexpression of lysozyme in the flies that expressed Ab 1-42 or AbPP-BACE1 did not change the soluble or insoluble level of Ab 1-42 in the two fly models (Fig. 4D,E). In order to investigate a potential interaction between Ab 1-42 and lysozyme, an immunoprecipitation assay was performed. Extract from AbPP-BACE1 flies with and without coexpression of lysozyme was run on a column with lysozyme antibodies bound to the resin. The urea-eluted samples were analysed for Ab 1-42 using the Meso Scale Discovery technique. The AbPP-BACE1-lysozyme extract had a significantly higher Ab 1-42 signal compared to the Ab 1-42 signal in the AbPP-BACE1 extract (Fig. 4C), which indicates an interaction between Ab 1-42 and lysozyme in the AbPP-BACE1-lysozyme flies. Taken together, these results demonstrate that the protective effect of lysozyme overexpression in the AD transgenic flies was not due to changed levels of Ab 1-42 , but might instead depend on interactions of Ab 1-42 and lysozyme in the AD fly eye.

Lysozyme rescues Ab Arc flies without changing the level of Ab Arc
To further investigate the beneficial effects of lysozyme on Ab-induced toxicity, a longevity assay was performed using a Drosophila model expressing Ab  with the Arctic mutation (Ab Arc ) with or without coexpressing human lysozyme in the fly CNS using the elav-Gal4 driver. As seen in Fig. 5A, there was a dosedependent rescue mediated by lysozyme where the median survival for Ab Arc flies was prolonged from 10 to 11 days for Ab Arc flies carrying one copy of lysozyme and from 10 to 12 days for Ab Arc flies carrying two copies of lysozyme. Flies expressing both lysozyme lines reduced the median survival for the control Gal4 expressing flies from 44 days to 41 and 39 days, line A and line B respectively. Next, the velocity and angle of movement were examined using the iFly technique [18]. To study the effect of lysozyme on toxicity induced by the Ab Arc peptide in adult flies, crosses were set up using the lower temperature 18°C, which has been demonstrated to significantly lower protein expression [19]. The flies were then moved to 29°C after eclosion to induce expression of Ab Arc as well as lysozyme. The velocity of the flies will decrease as they age or get sick. Shortly before the flies die, they become immobile and their velocity cannot be recorded; thus, a cut-off value of 4 mmÁs À1 was set as an indication of dysfunctional locomotor behaviour [10]. The Ab Arc flies reached below this value after 9 days and after 10 days, no movement could be recorded. For Ab Arc flies coexpressing one or two copies of lysozyme, the cut-off value was reached after 12 and 15 days respectively, revealing a dose-dependent rescue effect of lysozyme on Ab Arc toxicity where the dysfunctional locomotor behaviour of the Ab Arc flies was postponed 3 and 6 days (Fig. 5B). For control flies only expressing Gal4 and for flies expressing one or two copies of lysozyme, the cut-off value was reached after 33, 27 and 28 days, respectively, revealing a detrimental effect of lysozyme on the locomotor performance of the Gal4-expressing flies (Fig. 5B). The Ab Arc flies coexpressing one or two copies of lysozyme had a significantly higher velocity compared with the Ab Arc flies at day 9 (Fig. 5C).
The angle of movement increases as the flies age or get sick and a cut-off value of 80°was set in order to indicate dysfunction of the flies' locomotor behaviour [10]. In the angle of movement analysis, Ab Arc flies reached above 80°after 6 days, while Ab Arc flies with one or two copies of lysozyme both reached above this value after 10 days (Fig. 5D). Control Gal 4 flies and flies coexpressing one or two copies of lysozyme did not display this dysfunction, all genotypes reached above the cut-off value after 18 days. The Ab Arc flies coexpressing one or two copies of lysozyme had a significantly lower angle of movement compared with the Ab Arc flies at day 6 ( Fig. 5E). In order to investigate whether coexpression of one or two copies of lysozyme in the Ab Arc flies had any effect on the levels of soluble and insoluble Ab Arc , the level of Ab Arc was measured in the head of the flies using the Meso Scale Discovery technique after ageing the flies for 10 days at 29°C after eclosion. The soluble levels of Ab Arc were low and no significant differences were detected between the different fly variants (Fig. 5F). When analysing the insoluble levels of Ab Arc, significant levels were detected in all Ab Arc expressing flies compared to control Gal4 flies coexpressing two copies of lysozyme (Fig. 5F). No significant differences were detected between Ab Arc flies and flies coexpressing the Ab Arc peptide with one or two copies of lysozyme (Fig. 5F). Taken together, these data show that lysozyme had a dose-dependent rescue effect on the Ab Arc toxicity when lysozyme expression was directed to the CNS, lysozyme increased the median survival and decreased dysfunctional locomotor behaviour observed for the Ab Arc flies, without changing the levels of soluble or insoluble Ab Arc .

Discussion
Previously, it was demonstrated that lysozyme is associated with Ab-plaques in human brain tissue [10]. This study demonstrates that both mRNA and protein levels of lysozyme were upregulated in AD mouse models and in human AD cases and that lysozyme associated with Ab-plaques in mice brain tissue. In the AD transgenic mouse models, hippocampus and cortex, but not cerebellum, displayed increased mRNA expression of lysozyme compared to WT mice. A similar lysozyme mRNA profile was detected in humans, with increased lysozyme expression in visual cortex and prefrontal cortex in AD patients, and again, no apparent change in cerebellum. The spread of Ab pathology starts in the areas of cortex and hippocampus and does not affect the cerebellum until late in the disease in both AD mouse models and AD patients [20][21][22]. In this study, a similar pattern of lysozyme expression was revealed both in the AD mouse models and AD patients. The expression of lysozyme in the cortex and hippocampus of AD transgenic mice correlated well with Ab pathology and the Ab pathology had a spatial similarity with the increase of lysozyme expression. It was recently shown that exposure of neuroglia cells to oligomeric Ab causes an intracellular upregulation of lysozyme as well as an increased secretion of lysozyme [10]. Increased lysozyme protein levels were found in the cortex of PS1 and AbPP Swe transgenic mice and in the temporal cortex of human AD patients. An increased level of lysozyme was also detected in CSF from AD patients. This was previously seen in two cohorts of CSF of biochemical diagnosed AD patients [10]. The CSF cohort used in this study was from both clinical and biochemical diagnosed AD patients. It is likely that the increased mRNA expression in affected brain regions gave rise to increased protein expression and was responsible for the increased protein level of lysozyme seen in CSF from AD patients. Thus, these in vivo and cellular results suggest that Ab triggers an increase of lysozyme expression and secretion. The expression of lysozyme in tau transgenic mice did not correlate with the agerelated increase in tangle load, as demonstrated by the sigmoidal relationship between levels of Ab and lysozyme. Tau pathology was increased at 8 months in the tau transgenic mice, both in cortex and hippocampus, but the increase in lysozyme did not appear until after 18 months. These results agree with the previous findings in the AD transgenic mouse models that amyloid plaque pathology correlates with overexpression of several immune genes, whereas this correlation was not found for the tau transgenic mouse model [15]. It was previously shown that lysozyme is able to rescue a Drosophila AD model where Ab 1-42 and lysozyme were expressed and secreted, using a secretion tag, in the CNS of the flies [10]. In this study, the Ab 1-42 production was targeted to the retina of the flies by use of the gmr-Gal4 driver. The benefit of this driver is that the eye is a nonessential organ for fly survival, and products from transgenes that interfere with ommatidia development can easily be visualized in the fly eye as a rough eye phenotype. Moreover, in this study, we further investigated the effect of lysozyme on Ab toxicity by introducing a new AD model where the Ab peptide was generated by the processing of AbPP with b-secretase followed by c-secretase. This new AD fly model is more Fig. 4. Lysozyme protects transgenic AD flies from Ab-induced toxicity. (A) Images of rough eye phenotype obtained by scanning electron microscopy of two AD Drosophila models expressing Ab  or AbPP-BACE1 with or without lysozyme (Lys) and controls; only expressing Gal4, AbPP and Lys at the day of eclosion (1509 magnification). Scale bars = 50 lm. (B) Quantification of the abnormal ommatidia seen in A (n ≥ 4 of each genotype). All samples are related to their respective control. Significant differences were determined by one-way ANOVA followed by Tukey's test, **P ≤ 0.01, ***P ≤ 0.001. Meso Scale Discovery analyses performed on fly heads from flies at the day of eclosion to quantify the levels of (D) soluble Ab 1-42 (n = 4) and (E) insoluble Ab 1-42 (n = 4). All samples are compared to their respective control. Significant differences were determined by one-way ANOVA followed by Tukey's test, **P ≤ 0.01, ***P ≤ 0.001. (C) Immunoprecipitation assay of extract from AbPP-BACE1 flies with and without coexpression of lysozyme. Brain homogenates were immunoprecipitated with lysozyme antibodies bound to resin and the urea-eluted samples were analysed for Ab 1-42 using the Meso Scale Discovery technique. The assay was performed on 100 flies of each genotype. Significant differences were determined by Student's t-test, *P ≤ 0.05.

3515
The  physiological relevant and produces Ab of various lengths [17]. Flies that expressed Ab 1-42 and flies that coexpressed AbPP and BACE1 had defects in the ommatidia structure. This phenotype was completely abolished when lysozyme was coexpressed, resulting in fly eyes with well-ordered ommatidia similar to control flies. Thus, in both AD models, the eye phenotype was rescued by coexpression of lysozyme which demonstrates that lysozyme was able to exert its antitoxic effect, both when the Ab peptide was expressed directly from the transgene and when it was generated by AbPP processing. In addition, the rescue effect of lysozyme on the AbPP-BACE flies reveals possible interactions of lysozyme with several Ab species generated in these flies.
In addition, we report that lysozyme also had a dose-dependent protective effect on the Arctic mutation of Ab 1-42 , when expressed in the CNS of the flies. The Arctic mutation of Ab 1-42 leads to an increased propensity of Ab to form oligomeric species and at a faster rate [23]. One copy of lysozyme increased the median survival of flies expressing Ab Arc with 1 day, and two copies extended the median survival with 2 days. To achieve a more complete picture of the health of the flies, a locomotor assay was used to study the velocity and the angle of movement of flies expressing the Ab Arc peptide alone or together with one or two copies of lysozyme. Coexpression of both one and two copies of lysozyme demonstrated beneficial effects on both the velocity and angle of movement of the Ab Arc flies, which proves that lysozyme has the ability to decrease locomotor disability caused by the Ab Arc peptide.
The rescue effect of lysozyme on Ab Arc toxicity in the longevity and locomotor assays cannot be attributed to a general beneficial effect of the fly health by expressing lysozyme in the fly CNS as flies only expressing lysozyme showed a reduction in the median survival time compared to Gal4 control flies (by approximately 4 days). A detrimental effect was also detected in the velocity experiment for lysozyme expressing flies compared to the Gal4 flies, while no general effect of lysozyme was detected in the angle of movement analysis of Gal4 flies. Altogether, these results demonstrate that lysozyme has a protective effect on Ab-induced toxicity.
One possible explanation for the protective action of lysozyme could be that it increases the degradation of the Ab peptide in the flies. This was shown in a study where coexpression of lysozyme in Ab 1-42 transgenic Drosophila increased the survival of the flies and reduced the levels of Ab 1-42 [10]. However, in this study, there was no difference in the amount of either soluble or insoluble Ab 1-42 between flies that expressed Ab 1-42 or Ab 1-42 coexpressed with lysozyme, or AbPP-BACE1 flies and flies with AbPP-BACE1 and lysozyme. The discrepancies between these studies are that in the previous study, Ab  and lysozyme were expressed in the CNS but in this study, the expression was targeted to the eyes of the flies. The degradative capacity of Ab 1-42 in the brain and eye can be different and lysozyme can interact with Ab 1-42 in another fashion in the brain compared with the eye, which might explain the discrepancies between the results.
We have previously shown in vitro that lysozyme binds to monomeric Ab 1-42 and prevents the formation of toxic Ab 1-42 species by reducing the aggregation propensity of Ab 1-42 [10]. Other studies have shown that lysozyme also inhibits the aggregation of Ab 1-40 and Ab 17-42 by binding during the early stages of fibrillation, thereby reducing toxic species and cytotoxicity [11,12,24]. We have also reported that lysozyme colocalizes with Ab plaques in human AD brain and with Ab aggregates in Ab-overexpressing Drosophila flies [10]. In this study, we show that lysozyme stains plaques in AbPP mice which indicate that fibrillar Ab and lysozyme are in close association. To investigate if lysozyme was able to bind Ab in the gmr-Gal4-derived flies, immunoprecipitation capturing lysozyme was performed on fly head homogenates followed by Meso Scale Discovery, where Ab was detected with the 6E10 antibody. Indeed, there was a significant increase of Ab 1-42 in the homogenates from AbPP-BACE1-lysozyme flies compared with the AbPP-BACE1 flies. This demonstrates that also in this Drosophila model lysozyme was able to interact Ab with the Arctic mutation (Ab Arc ) in the CNS, in the absence or presence of lysozyme expressed as one copy or as two copies compared to control flies (n = 100). Kaplan-Meier graph shows per cent survival against age of flies in days after eclosion. Significant differences were determined by log-rank analysis, *P ≤ 0.05 and ***P ≤ 0.001. The fly behaviour was analysed for control flies and Ab Arc flies with or without coexpression of one or two copies of lysozyme by (B) velocity measurement, (C) velocity measurement at day 9, (D) angle of movement analysis and (E) angle of movement analysis at day 6. Significant differences were determined by one-way ANOVA followed by Tukey's test, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. Error bars represent mean AE SD, n = 30. (F) Meso Scale Discovery analysis performed on the fly heads of control, Ab Arc , Ab Arc -Lys and Ab Arc -Lys-Lys at day 10 after eclosion to quantify the levels of soluble and insoluble Ab Arc . Significant differences were determined by one-way ANOVA followed by Tukey's test, ***P ≤ 0.001. Error bars represent mean AE SD, n = 4. with Ab and thus likely rescues the degeneration of the ommatidia in the fly eye by preventing the formation of cytotoxic Ab species.
When quantifying the soluble and insoluble levels of Ab Arc in the fly CNS, only insoluble Ab species were detected and no significant differences could be observed between the insoluble levels in flies, solely expressing Ab Arc , and in flies where Ab Arc was coexpressed with one or two copies of lysozyme. Hence, the protective properties of lysozyme in the longevity and locomotor analyses of the Ab Arc flies must be unrelated to degradation of Ab Arc . A possible explanation for the antitoxic effect of lysozyme in the Ab Arc flies could be due to interactions between the insoluble Ab Arc species and lysozyme reducing the toxicity of these species. The Arctic mutation causes the Ab 1-42 peptide to be more aggregation prone and therefore less prone to be degraded. This might explain why lysozyme overexpression in the CNS of Ab Arc flies was unable to cause an increased degradation of Ab, which earlier was suggested for lysozyme overexpression in the CNS of Ab 1-42 flies [10].
In conclusion, we show that the expression of lysozyme was increased in the brains of both transgenic AD mice and humans with AD. Furthermore, we demonstrate favourable effects of lysozyme when expressed in different Drosophila models of AD. In flies that expressed Ab  or AbPP together with BACE1 in the eyes, the rough eye phenotype indicative of toxicity was completely rescued by coexpression of lysozyme. In Drosophila flies that express the toxic Ab Arc , lysozyme both increased the fly survival dose dependently and decreased dysfunction in locomotor behaviour. We propose that this is due to binding of lysozyme to toxic species of Ab which prevent these from exerting their toxic effect. Thus, the upregulated of lysozyme expression in AD patients and transgenic AD mice might be a rescue response towards Ab toxicity. These results highlight the possibility of lysozyme as a potential therapeutic target for AD.

Gene expression studies
The database (www.mouseac.org) which includes microarray data from amyloid transgenic mice carrying the AbPP Swe mutation and/or PSEN1 (PSEN1: M146V) either heterozygous or homozygous, and tau transgenic mice (TAU: P301L) [15], was used to study lysozyme mRNA expression and pathology during ageing (GEO accession number GSE64398). In addition, a microarray data set containing information about human mRNA expression was used to study the lysozyme expression in 101 nondemented healthy controls and 129 AD patients (GEO accession number GSE44772) [16].
Protein extraction from brain tissue Human brain tissue were obtained from the Sydney Brain Bank at Neuroscience Research Australia and the New South Wales Tissue Resource Centre at the University of Sydney and characterized according to established neuropathological criteria [25]. Informed consent for the collection of material was obtained prior to death and tissue use was approved by the University of New South Wales Human Research Ethics Committee. Brain biopsies from 34 cases (Braak stages 0-VI) from temporal cortex were frozen with a post-mortem delay of maximum 64 h. The tissues were homogenized in 4 lLÁmg À1 sample of cold TBS [50 mM TRIS, 125 mM NaCl, 5 mM EDTA, protease inhibitor (Roche, Basel, Switzerland)] using a motorized pestle prior to centrifugation at 20 800 g at 4°C for 1.5 h.
The APP-PS1 AD mouse model expressing chimeric mouse/human APP695swe/Swedish mutations (K595N/ M596L) and mutant human PS1 (PS1/ΔE9) was obtained from the Jackson Laboratory (Bar Harbor, ME, USA; Strain name, B6.Cg-Tg AbPPswePSEN1dE9) 85Dbo/J; Stock #005864). Animal ethics approval was from the University of Wollongong Animal Ethics Committee (AE11/ 03). The mice were aged for 12 months without any intervention; food and water were available ad libitum. In brief, 12month-old APP-PS1 and age-matched wild-type mice were euthanized by CO 2 asphyxiation to ensure that the mice did not suffer unnecessarily, and transcardially perfused with ice-cold phosphate-buffered saline (PBS). The right hemisphere from the APP-PS1 and age-matched wild-type mice were snap frozen and stored at À80°C. Cortex was dissected and homogenized in 10 volumes of 140 mM NaCl, 3 mM KCl, 25 mM Tris (pH 7.4), containing 1% Nonidet P-40 and Roche complete protease inhibitors using a Precellys 24 homogenizer (2 9 30 s, 6000 g). The supernatants were saved and the protein concentration was determined using a DC protein assay (Bio-Rad, Hempstead, UK).

CSF cohort
Deidentified and archived CSF samples, that were both biochemically and clinically diagnosed with AD (n = 25), were age-and gender-matched with controls with normal levels of the CSF AD biomarkers P-tau 181P , T-tau and Ab 1-42 (n = 25). The samples were provided by the Clinical Neurochemistry Laboratory, Sahlgrenska University Hospital/M€ olndal, Sweden. The study was approved by the Ethical committee at the University of Gothenburg. For more information regarding CSF handling and P-tau 181P , T-tau and Ab 1-42 analysis, see previously described methods [26].

Western blotting
Protein separation was performed as described previously [26]. Forty microgram of total protein was mixed with loading buffer and blotted onto a nitrocellulose membrane using an iBlot Dry Blotting System (Invitrogen, Paisley, UK) and incubated with polyclonal rabbit anti-human lysozyme antibody (1 : 2000) (Dako, Glostrup, Denmark) overnight in 4°C, followed by horseradish peroxidase (HRP)-linked goat anti-rabbit antibody (Dako) for 1 h at room temperature. Mouse anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH; 1 : 20 000; Novus Biologicals, Littletown, CO, USA) was used as loading control. To correct for differing protein levels between blots, all samples were normalized to a reference sample, which was loaded onto each gel.

Immunohistochemistry
C57Bl/6 transgenic AbPP Swe mice (kindly provided by Professor Lars Nilsson at Oslo University) were housed in a pathogen-free environment on a 12-h light/dark cycle. The mice were aged for 15 months without any intervention; food and water were available ad libitum. The experimental procedures were performed in accordance with the Animal Care and Use Ethical Committee at the Link€ oping University (ethical registration number 84-12) which follow the directives 2010/63/EU of the European Parliament and of the Council of 22 September 2010 on the protection of animals used for scientific purposes. The mice were euthanized by CO 2 asphyxiation to ensure that the mice did not suffer unnecessarily, and transcardially perfused with saline solution (NaCl, 0.9%) before dissection of the brain. The brain tissue were stored in formalin, dehydrated in sucrose, sectioned and frozen. Immunohistochemistry was performed according to the manufacturer's protocol (Biocare Medical, Concord, CA, USA). Ab plaques were immunostained using the 6E10 antibody (1 : 100; Signet Laboratories/Covance, Dedham, MA, USA), then colour was developed using diaminobenzidine (DAB; Biocare Medical) and lysozyme was immunostained with an anti-lysozyme antibody (1 : 100; Dako) and colour was developed with Vina Green TM Chromogen Kit (Biocare Medical). To stop crossreaction, the brain slices were incubated with denaturing solution (Biocare Medical) for 5 min before double staining with lysozyme antibody.

Drosophila stocks
To direct tissue-specific protein expression in UAS transgenic Drosophila melanogaster, the Gal4/UAS system was used [27]. In the SEM analysis, the tissue-specific gmr-Gal4 driver strain was used to direct protein expression to the photoreceptors present in the retina of the flies [28]. Control w 1118 flies (only expressing Gal4), AbPP 695 (AbPP) and BACE1-expressing flies were purchased from Bloomington stock centre. Signal-peptide-Ab 1-42 flies and signal-peptide-Ab Arc (E22G) were kindly provided by Professor Damian Crowther [29]. Lysozyme-expressing flies were constructed as described previously [30]. To generate flies coexpressing Ab  or AbPP-BACE1 together with lysozyme or flies coexpressing Ab Arc with either one or two copies of WT lysozyme, cross-breeds were set up in multiple steps with 12:12 h light:dark cycles and 60% humidity. In the longevity and locomotor analyses, the elav-Gal4 driver line was used to direct protein expression to the fly CNS. To generate X Elav /Y;Ab Arc /Ife;TM6b/+, a cross was set up using X Elav /X Elav ;CyO/Ife;TM6b/mkrs and X/Y;Ab Arc /Ab Arc ;+/+. In order to achieve coexpression of Ab Arc with one or two copies of lysozyme, males generated in the cross previously described were crossed with X/X;Lys lineB /Lys lineB ;+/+ or X/ X;Lys lineB /Lys lineB ;Lys lineA /TM6b. All gene constructs of lysozyme and Ab Arc where cloned into the Gal4-responsive pUAST expression vector to generate UAS-lysozyme or UAS-arctic transgene. These constructs were injected into the same background W 1118 flies [29,30].

SEM of eye phenotype
Cross-breeds were set up at 25°C. At the day of eclosion, flies were euthanized with ether before they were mounted onto 12-mm aluminium specimen stubs with an adhesive tape and air-dried for 24 h. The samples were sputter coated with 15 nm of platinum and stored in a vacuum desiccator prior to analysis using SEM (JEOL JSM-6320F). The eyes were scanned at 1509 magnification at an accelerating voltage of 10 kV. Images were recorded using the SemA for 5.21 digitizer system (Insinooritoimisto Rimppi Oy). In a blinded set-up, the images were printed and assigned a square of approximately 100 ommatidia in the centre of the eye. All abnormal ommatidia were quantified within this square and further related to all ommatidia. Scale Discovery, Rockville, ML, USA). The samples were stored at À80°C. To account for differences in the protein extraction step, the total amount of protein extracted was quantified using the Bio-Rad DC Protein Assay Kit II (500-0112, Bio-Rad).

Quantification of Ab 1-42 by Meso Scale Discovery analysis
The wells in a multi-spot 96-well V-PLEX human Ab 1-42 (K151LBE-1, Meso Scale Discovery) were blocked using 150-lL Diluent 35 for 1 h at RT with gentle agitation. After blocking, 50 lL of each prepared protein sample was added to the wells in duplicates (incubated 1 h, RT, gentle agitation). After sample incubation, the wells were washed in 3 9 150 lL PBS-T before adding 25 lL of the detection antibody (509 Sulfo tag 6E10, Meso Scale Discovery) (1 h, RT, gentle agitation). The wells were then washed again (3 9 150 lL PBS-T) before adding 150 lL 29 read buffer (10 min incubation, RT, no agitation). The plate was analysed using a SECTOR Imager 2400 instrument (Meso Scale Discovery).

Sample preparation and quantification of Ab Arc by Meso Scale Discovery
Fly crosses were set up at 18°C and at the day of eclosion moved to 29°C where they were aged for 10 days. After ageing, 10 fly heads of each genotype were placed in Eppendorf tubes and stored at -80°C until use. The Meso Scale Discovery analysis was performed as previously described by Caesar et al. [31]. To account for any differences in the protein extraction step, the total amount of protein extracted for each sample was quantified using the Bio-Rad DC Protein Assay Kit II (500-0112, Bio-Rad).

Longevity assay
Fly crosses using the elav-Gal4 driver were set up at 29°C and maintained at 29°C (60% humidity, 12:12 h dark:light cycle) after eclosion. A set of 100 female offspring for each genotype was collected at the day of eclosion. The flies were then divided into groups of approximately 10 flies and placed in vials containing agar food (20 g agar, 20 g sugarÁL À1 dH 2 O) and yeast paste (dry baker's yeast mixed with water). Every 2-3 day, the flies were transferred to vials containing fresh food and the number of living flies was counted. This was repeated until all flies had died. Kaplan-Meier survival curves were generated for lifespan assessment.

Locomotor assay
Fly crosses were set up at 18°C and at the day of eclosion, moved to 29°C. Sets of 30 female flies of each genotype were collected and divided into groups of 10 and placed in vials containing fly food. To analyse the flies' locomotor behaviour, the flies were filmed during 90 s and tapped to the bottom of the vial every 30 s to reactivate locomotor behaviour. This was carried out every day until all Ab Arcexpressing flies had stopped to move; the Gal4-and lysozyme-expressing flies were analysed every other/third day until no locomotor behaviour could be detected for these fly variants. The videos were processed and analysed using the IFLY software [18], which calculated the velocities and angles of movement generated by the flies in each recorded video clip. The locomotor measurements were carried out as described previously [17].

Immunoprecipitation
Lysozyme-specific capture resin was prepared as described earlier [30] with CNBr-activated Sepharose 4B resin (GE Healthcare, Pittsburgh, PA, USA) and 0.5 mgÁmL À1 polyclonal rabbit anti-human lysozyme (Dako; A0099). One hundred microlitre of the anti-lysozyme conjugated medium slurry was placed in a Micro Bio-Spin Chromatography Column (Bio-Rad) and the column was equilibrated with SuperBlock TBS Blocking Buffer (Thermo Scientific Pierce, Rockville, IL, USA) supplemented with a protease inhibitor (TBS-PI). Heads from AbPP flies, AbPP-BACE flies and AbPP-BACE flies with lysozyme were homogenized with a pestle in 200 lL TBS-PI and centrifuged for 1 min at 16 000 g. One hundred flies of each genotype were used. The supernatant was centrifuged a second time before incubation in the column for 5 min and centrifuged 160 g for 1 min. The resin was then washed four times with TBS-PI before lysozyme was eluted with 6 M urea. The concentration of Ab 1-42 was measured using the Meso Scale Discovery technique, as described above.

Statistical analysis
All statistical analyses and graphs were performed and drawn using GRAPHPAD PRISM software v. 7 (GraphPad Software, La Jolla, CA, USA). Two-way ANOVA followed by Tukey's post hoc test was used to test for significant differences between multiple groups of mice. Correlation analysis was performed using the Pearson correlation coefficient. One-way ANOVA followed by Tukey's test was used for quantification of abnormal ommatidia and to test for differences in protein levels analysed with the Meso Scale Discovery technique. Student's t-test or Mann-Whitney U test was used to test for significant differences between two groups. For the lifespan assay, Kaplan-Meier survival curves were generated and log-rank statistical analysis (Mantel-Cox) was performed. Statistical significance was defined for P-values of < 0.05 (*), < 0.01 (**) and < 0.001 (***).