Abstract

Background

   Haploinsufficiency of the progranulin (PGRN) protein is a leading cause
   of frontotemporal lobar degeneration (FTLD). Mouse models have been
   developed to study PGRN functions. However, PGRN deficiency in the
   commonly used C57BL/6 mouse strain background leads to very mild
   phenotypes, and pathways regulating PGRN deficiency phenotypes remain
   to be elucidated.

Methods

   We generated PGRN-deficient mice in the FVB/N background and compared
   PGRN deficiency phenotypes between C57BL/6 and FVB/N backgrounds via
   immunostaining, western blot, RNA-seq, and proteomics approaches. We
   demonstrated a novel pathway in modifying PGRN deficiency phenotypes
   using inhibitor treatment and AAV-mediated overexpression in mouse
   models.

Results

   We report that PGRN loss in the FVB/N mouse strain results in earlier
   onset and stronger FTLD-related and lysosome-related phenotypes. We
   found that PGRN interacts with sPLA2-IIA, a member of the secreted
   phospholipase A2 (sPLA2) family member and a key regulator of
   inflammation, that is expressed in FVB/N but not C57BL/6 background.
   sPLA2-IIA inhibition rescues PGRN deficiency phenotypes, while
   sPLA2-IIA overexpression drives enhanced gliosis and lipofuscin
   accumulation in PGRN-deficient mice. Additionally, RNA-seq and
   proteomics analysis revealed that mitochondrial pathways are
   upregulated in the PGRN-deficient C57BL/6 mice but not in the FVB/N
   mice.

Conclusions

   Our studies establish a better mouse model for FTLD-GRN and uncover
   novel pathways modifying PGRN deficiency phenotypes.

Supplementary Information

   The online version contains supplementary material available at
   10.1186/s13024-025-00863-8.

   Keywords: Progranulin, Inflammation, Lysosome, Mouse strain background,
   sPLA2-IIA, Mitochondria

Introduction

   Progranulin (PGRN), encoded by the granulin (GRN) gene in humans, has
   emerged as a key player in brain aging and neurodegenerative diseases
   [[50]1, [51]2]. Mutations in the GRN gene, resulting in PGRN
   haploinsufficiency, are one of the major causes of frontotemporal lobar
   degeneration (FTLD) with TDP-43 positive inclusions [[52]3–[53]5].
   Complete loss of PGRN in humans causes neuronal ceroid lipofuscinosis
   (NCL) [[54]6–[55]8], a group of lysosomal storage diseases (LSDs). GRN
   was also identified as one of the two main determinants of differential
   aging in the cerebral cortex [[56]9] and as one of the five risk
   factors for limbic-predominant age-related TDP-43 encephalopathy (LATE)
   [[57]10]. Polymorphisms in the GRN gene also contribute to the risk of
   Alzheimer’s disease (AD) [[58]11–[59]14], and GRN is closely linked to
   Parkinson’s disease (PD) [[60]15, [61]16].

   PGRN is an evolutionarily conserved glycoprotein comprised of 7.5
   granulin repeats with critical roles in inflammatory responses
   [[62]17–[63]20]. PGRN is highly expressed in microglia, the resident
   immune cell in the brain, and regulates many aspects of microglial
   behavior [[64]20]. PGRN deficiency results in exacerbated microglial
   activation, enhanced phagocytosis, increased levels of complement
   factors, and excessive synaptic pruning, leading to the disruption of
   brain circuits [[65]21]. At the cellular level, PGRN is either secreted
   to the extracellular space or sorted into the lysosomal compartment
   within the cell [[66]2, [67]22]. PGRN is processed into granulin
   peptides in the lysosome and is critical for proper lysosomal
   functions, especially during aging [[68]2, [69]22]. Despite these
   studies, the precise mechanisms by which PGRN regulates brain
   inflammation and lysosomal functions remain elusive.

   It is well established that the genetic background of mouse strains
   influences disease-associated phenotypes [[70]23–[71]26]. Various mouse
   strains, such as C57BL/6 and BALB/c, have been used to explore the
   genetic and phenotypic variability in the manifestation of
   neurodegenerative diseases, including multiple LSDs and AD
   [[72]27–[73]29]. Commonly used PGRN knockout (Grn^−/−) mouse lines in
   the C57BL/6 (hereafter B6) background are known to develop gliosis and
   lysosomal abnormalities in an age-dependent manner, with mild glia
   activation and lipofuscin accumulation typically observed by 7 months
   of age [[74]30]. In addition, no detectable or very mild TDP-43
   pathology is observed in aged Grn^−/− mice in this background [[75]30,
   [76]31]. In this study, we report that PGRN deficiency in the FVB/N
   (hereafter FVB) background leads to earlier and more severe phenotypes
   compared to the B6 background, including elevated microglia and
   astrocytes activation, lysosomal defects, and TDP-43 pathology. To
   dissect the mechanisms behind the strain background differences, we
   investigate the role of sPLA2-IIA, a secreted phospholipase found to
   interact with PGRN in a proteomic screen. Interestingly, sPLA-IIA is
   expressed in FVB but not B6 strains. We showed that sPLA2-IIA
   activities and levels are downregulated by PGRN. Furthermore, sPLA2
   inhibitors rescue PGRN deficiency phenotypes in the FVB mice, and
   AAV-mediated sPLA2-IIA overexpression leads to enhanced phenotypes in
   Grn^−/− mice in the B6 background. To further explore the mechanisms
   behind the strain background differences affecting phenotypes
   associated with PGRN deficiency, we performed RNA-seq and proteomics
   analyses and found pronounced transcriptome and proteome differences
   between the FVB and B6 strains. Intriguingly, the mitochondrial
   pathways are differentially regulated by PGRN deficiency in the two
   backgrounds, which might affect PGRN deficiency phenotypes. Taken
   together, our studies establish a better model to study PGRN function
   in mice and identify novel pathways regulating PGRN deficiency
   phenotypes.

Materials and methods

DNA and plasmids

   Human Pla2g2a cDNA in the pDONR221 vector was obtained from DNASU.
   Human Pla2g2a-Myc-His construct was obtained through a gateway reaction
   with the pDONR221-Pla2g2a entry vector and an engineered destination
   vector with CMV promoter and Myc-His tag (h-sPLA2-IIA-myc). Human PGRN
   construct was previously described [[77]32, [78]33]. Mouse
   Pla2g2a-Myc-DDK construct was obtained from Origene (m-sPLA2-IIA-myc).
   Mouse PGRN construct was cloned into pSecTag2B vector (Invitrogen) with
   an N-terminal FLAG-His tag (Flag-mPGRN).

Primary antibodies and reagents

   The following antibodies were used in this study: mouse anti-Myc (9E10)
   (Sigma, 1:1000 for western blot), mouse anti-GAPDH (Proteintech Group,
   60004-1-Ig, 1:10000 for western blot), mouse anti-actin (Proteintech
   Group, 66009-1-Ig, 1:5000 for western blot), rat anti-mouse LAMP1
   (1D4B) (BD Biosciences, 553793, 1:300 for immunostaining), sheep
   anti-mouse PGRN (R&D systems, AF2557, 1:1000 for western blot and 1:100
   for immunostaining), goat anti-CathD (R&D systems, AF1029, 1:1000 for
   werstern blot and 1:100 for immunostaining), sheep anti-mouse sPLA2-IIA
   (ThermoFisher Scientific, PA5-47672, 1:1000 for western blot), mouse
   anti-NeuN (Millipore, MAB377, 1:500 for immunostaining), rabbit
   anti-IBA-1 (Wako, 01919741, 1:500 for immunostaining), rat anti-CD68
   (Bio-Rad, MCA1957, 1:300 for immunostaining), mouse anti-GFAP (Cell
   signaling, 3670 S, 1:3000 for western blot and 1:500 for
   immunostaining), mouse anti-galectin-3 (BioLegend, 126702, 1:1000 for
   western blot and 1:300 for immunostaining), goat anti-GPNMB (R&D
   Systems, AF2330, 1:300 for immunistaining), rabbit anti-TDP43
   (Proteintech Group, 12892-1-AP, 1:5000 for western blot), and rabbit
   anti-phospho-TDP-43 (Ser409/410) (Proteintech group, 80007-1-RR, 1:5000
   for western blot). Myc-Trap (ChromoTek) and GFP-trap (Proteintech
   Group) were used for anti-myc and anti-GFP immunoprecipitations,
   respectively.

   The following reagents were also used in the study: Dulbecco’s modified
   Eagle’s medium (DMEM) (Cellgro, 10-017-CV), 0.25% Trypsin (Corning,
   25-053-CI), [79]LY333013 (Chemietek, CT-VAREM), KH064 (Millipore Sigma,
   S3319), Odyssey blocking buffer (LI-COR Biosciences, 927-40000),
   protease inhibitor (Roche, 05056489001), Pierce BCA Protein Assay Kit
   (Thermo Scientific, 23225), O.C.T compound (Electron Microscopy
   Sciences, 62550-01), mouse sPLA2 activity assay Kit (Cayman Chemical,
   765001).

Mouse strains and treatments

   Grn^−/− mice in the B6 background [[80]31] mice were obtained from
   Jackson Labs and housed in the Weill Hall animal facility. To generate
   the Grn^−/− mice in the FVB background, male B6 Grn^−/− mice
   (Pla2g2a^−/−Grn^−/−) were mated to female WT FVB (Pla2g2a^+/+Grn^+/+)
   mice to generate Pla2g2a^+/−Grn^+/− mice. Pla2g2a^+/−Grn^+/− mice were
   bred with each other to generate Pla2g2a^+/+Grn^−/− mice in the FVB/B6
   background. An additional 6 backcrosses with female wild-type FVB were
   made to generate Grn^+/− mice in a pure FVB background. To generate
   Grn^−/− mice with wild type Pla2g2a expression on a B6 background, FVB
   Grn^−/− mice (Pla2g2a^+/+Grn^−/−) were mated to female B6 Grn^−/− mice
   (Pla2g2a^−/−Grn^−/−) to generate Pla2g2a^+/−Grn^−/− mice.
   Pla2g2a^+/−Grn^−/− mice were bred with B6 mice for an additional 6
   backcrosses to get these mice with a pure B6 background. Pla2g2a
   genotyping was performed using the following primers: Forward common,
   5’- CAGAGCTGACAGCATGAAGGTCCTC-3’; Reverse WT,
   5’-TCTGTGGCATCCTTGGGGGAT-3’; Reverse pla2g2a knockout, 5’-
   CTGTGGCATCCTTGGGGGAA-3’.

   For inhibitor treatment, 5-week-old Pla2g2a^+/+Grn^−/− mice were fed
   with 0.3% DMSO in corn oil, [81]LY333013 (10 mg/kg, dissolved in DMSO
   and diluted in corn oil) or KH064 (10 mg/kg, dissolved in DMSO and
   diluted in corn oil) via oral gavage once per day for 3 consecutive
   weeks. Both male and female mice were included in each group. The age
   of the mice used for each experiment is described in the figure legend.
   All the mice were housed in the Weill Hall animal facility at Cornell.
   All animal procedures have been approved by the Institutional Animal
   Care and Use Committee (IACUC) at Cornell.

Adeno-associated virus (AAV) production and injection

   The AAV-Pla2g2a transgene (pAAV-mPla2g2a) was generated by cloning
   mouse pla2g2a cDNA into a single-stranded AAV vector containing a
   chicken β-actin (CBA) promoter and a bovine growth hormone
   polyadenylation (BGH-PolyA) signal. A Kozak sequence was added for
   efficient initiation of translation, and the woodchuck hepatitis virus
   posttranscriptional regulatory element (WPRE) was incorporated to
   enhance sPLA2-IIA expression. The ssAAV-Pla2g2a was packaged into AAV9
   capsids, and the virus was purified and titrated using a q-PCR-based
   approach by VectorBuilder. For AAV injection in mice, newborn WT and
   Grn^−/− C57BL/6 mouse pups (P0) were intravenously injected with 4 µL
   of AAV9-mPla2g2a (1.47 × 10^13 GC/mL) or the control buffer (10 mM
   Phosphate, 138 mM NaCl, 2.7 mM KCl, 0.001% Pluronic-68) via
   retro-orbital route as described previously [[82]34]. Mice were
   collected at 2 months old.

Cell culture and DNA transfection

   HEK293T cells were maintained in Dulbecco’s Modified Eagle’s Medium
   (Cellgro) supplemented with 10% fetal bovine serum (Sigma) in a
   humidified incubator at 37ºC with 5% CO2. For transient overexpression,
   HEK293T Cells were transfected with polyethyleneimine [[83]35].
   Briefly, plasmids were diluted in DMEM, mixed with polyethyleneimine
   (PEI), and incubated for 15 min at room temperature. The mixture was
   then added to cells, followed by culture in DMEM supplemented with 10%
   FBS for 3 days.

Tissue lysate Preparation

   Mice were perfused with PBS, and tissues were dissected and snap-frozen
   with liquid nitrogen and kept at -80 °C. On the day of the experiment,
   frozen tissues were thawed and homogenized on ice with bead homogenizer
   (Moni International) in ice-cold RIPA buffer (150 mM NaCl, 50 mM
   Tris-HCl [pH 8.0], 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS)
   with 1 mM PMSF, and 1x protease inhibitors (Roche). After
   centrifugation at 14,000 × g for 15 min at 4℃, supernatants were
   collected for analysis. The insoluble pellets were washed with RIPA
   buffer and extracted in 2× v/w of Urea buffer (7 M Urea, 2 M Thiourea,
   4% CHAPS, 30 mM Tris, pH 8.5). After sonication, samples were
   centrifuged at 200,000 g at 24 °C for 1 h, and the supernatant was
   collected as the Urea-soluble fraction.

Protein purification

   His-PGRN and sPLA2-IIA proteins were purified from the serum-free
   culture medium of transfected HEK293T cells using cobalt beads. Protein
   was eluted with imidazole. Purified proteins were further concentrated
   and changed to PBS buffer using the Centricon device (Millipore).

SILAC proteomics analysis

   HEPG2 cells were grown for a minimum of five generations in DMEM with
   10% dialyzed FBS (Sigma) supplemented with either 12C14N or 13C15N
   Arginine and Lysine. To search for PGRN interactors, the heavy HEPG2
   cells were transfected in two 15 cm dishes with GFP-PGRN expression
   constructs, while the light HEPG2 cells were transfected with pEGFP-C1
   as a control. Two days after transfection, cells were lysed in 50mM
   Tris pH8.0, 150 mM NaCl, 1% Triton, and 0.1% deoxycholic acid with
   protease inhibitors (Roche). The lysates were subjected to anti-GFP
   immunoprecipitation using GFP-Trap beads (ChromoTek). The presence of
   GFP and GFP-PGRN in immunoprecipitated samples was confirmed by
   SDS-PAGE and silver staining. Samples were then mixed and boiled for
   5 min with 1% DTT followed by alkylation by treating samples with a
   final concentration of 28 mM iodoacetamide. Proteins were precipitated
   on ice for 30 min with a mixture of 50% acetone/49.9% ethanol/0.1%
   acetic acid. The protein pellet was washed with the same solution and
   resolubilized in 8 M urea/50 mM Tris pH 8.0 followed by dilution with
   three volumes of 50 mM Tris pH 8.0/150 mM NaCl. Proteins were digested
   overnight on a nutator at 37ºC with 1 µg mass spectrometry grade
   Trypsin Gold (Promega). The resulting peptide sample was acidified at a
   final concentration of 0.25% formic acid and 0.25% trifluoroacetic acid
   (TFA) and desalted using a 50 mg Sep-Pak C18 cartridge (Waters).
   Samples were eluted with 80% acetonitrile (ACN)/0.1% acetic acid into
   silanized vials (National Scientific) and evaporated using a SpeedVac.
   Samples were reconstituted in 85 µL of 700:300:1 ACN: Water: FA and
   fractionated via hydrophilic interaction liquid chromatography (HILIC)
   on an Ultimate 300 LC (Dionex) using a TSKGel Amide-80 column (Tosoh
   bioscience). Each fraction was evaporated with a SpeedVac and
   resuspended in 0.1% TFA. Samples were subjected to LC-MS/MS analysis
   using a Q-Exactive HF mass spectrometer (Thermo Fisher Scientific), and
   data were analyzed using the SORCERER 2 system (Sage-N Research).

Immunoprecipitation and Western blot analysis

   Cells were lysed in 50 mM Tris pH 8.0, 150 mM NaCl, 1% Triton, and 0.1%
   deoxycholic acid with protease inhibitors (Roche). The lysates were
   subject to anti-GFP immunoprecipitation using GFP-Trap beads or
   anti-Myc immunoprecipitation using Myc-Trap beads (ChromoTek). The
   beads were washed with 50 mM Tris pH 8.0, 150 mM NaCl, 1% Triton after
   3–4 h of incubation. Western Blot was visualized using the
   Licor-Odyssey system as described [[84]36].

Immunofluorescence staining and image acquisition

   For mouse brain section staining, mice were perfused with cold PBS, and
   tissues were post-fixed with 4% paraformaldehyde. After dehydration in
   30% sucrose buffer, tissues were embedded in the O.C.T compound
   (Electron Microscopy Sciences). 18-µm-thick brain sections were cut
   with a cryotome. Antigen retrieval was performed by microwaving in
   citrate buffer (pH 6.0) for 20 min. Tissue sections were blocked and
   permeabilized with 0.1% saponin in Odyssey blocking buffer before
   incubating with primary antibodies overnight at 4℃. The next day,
   sections were washed 3 times with cold PBS, followed by incubation with
   secondary fluorescent antibodies and Hoechst at room temperature for
   two hours. The slides were then mounted using a mounting medium (Vector
   Laboratories). Images were acquired on a CSU-X spinning disc confocal
   microscope (Intelligent Imaging Innovations) with an HQ2 CCD camera
   (Photometrics) using 63x or 100x objectives, and ten to twenty
   different random images were captured. Lower magnification images were
   captured by 10x or 20x objectives on a Leica DMi8 inverted microscope,
   and three to five images were captured from each sample.

sPLA2 activity assays

   sPLA2 activities were analyzed using a mouse sPLA2 activity assay kit
   (Cayman Chemical) according to the manufacturer’s instructions. Samples
   were prepared in a 96-well plate and run at 414 nm every minute for
   10 min. The activities of recombinant sPLA2-IIA protein (250 ng/ml) in
   the absence or presence of recombinant PGRN or BSA (1250 ng/ml or 2500
   ng/ml) were determined.

Lipid measurement

   To measure prostaglandin and free fatty acids levels, lipids were
   extracted from 50 mg of the frozen cortices and 80 µL of serum using
   the Bligh-Dyer protocol [[85]37]. Briefly, 450 µL methanol/chloroform
   (2:1, v/v) was added to pulverized frozen brain tissues or serum, and
   the samples were vortexed. 120 µL of water was then added to each
   tissue suspension, and the samples were vortexed for 2 min, followed by
   centrifugation at 13,000 rpm for 2 min. 150 µL chloroform and water
   were then added sequentially, and the samples were vortexed and
   centrifugated as above after adding each reagent. The lower organic
   layers were collected as lipid extracts. The lipid extracts were
   evaporated under a stream of nitrogen, dissolved in 100 µL
   chloroform/methanol/water (73:23:3, v/v), and then analyzed by mass
   spectrometry (6230 Time-of-Flight, Agilent) coupled to high-performance
   liquid chromatography (1260 Infinity, Agilent). Samples were separated
   using normal phase with a luna silica column (3 μm, 50 × 2 mm,
   Phenomenex) and a binary gradient elution system where solvent A was
   chloroform/methanol/ammonium hydroxide (85:15:0.5, v/v) and solvent B
   was chloroform: methanol: water: ammonium hydroxide (60:34:5:0.5, v/v)
   under the flow rate of 0.5 mL/min. The following gradient time program
   (time [%A/%B]) was used: 0 min (100/0)-0.5 min (100/0)-9 min
   (0/100)-12 min (0/100)-15 min (100/0)-18.24 min (100/0). Compounds were
   detected using a Dual Agilent Jet Stream Electrospray Ionization source
   operated in negative mode with the following parameters: gas
   temperature, 325 ºC; drying gas, 10 L/min; nebulizer, 20 psig; sheath
   gas temperature, 400ºC; sheath gas flow, 12 L/min; VCap, 4000 V; Nozzle
   Voltage, 2000 V; fragmentor, 175 V m/z; mass range, 100–1500;
   acquisition rate, 1000 ms/spectrum. Data were acquired and converted
   to.mzdata files by MassHunter Workstation Software (Version B.08.00,
   Agilent) for the subsequent analysis. Signals within the range of
   mz = 351.225 ± 0.05 were integrated to generate the extracted ion
   chromatograms, and chromatographic peaks were integrated using Python
   3.10.4 available at [86]http://www.python.org. Peaks were only selected
   when their intensity exceeded three times the baseline noise.
   Prostaglandins were quantified using a standard curve generated from
   serial dilutions of standard PGE2 (P1884, TCI), which is estimated to
   be the most abundant prostaglandin with a de-protonated mass of 351.225
   in the serum sample. The final prostaglandin concentrations were
   standardized per milliliter of serum or per milligram of brain tissue.
   For C18:1, C18:2, and C18:3 fatty acids measurement, the total peak
   area identified for each lipid species was measured, and the relative
   changes were analyzed.

RNA-seq analysis

   Total RNA was extracted from cortex regions using Trizol (Thermo
   Scientific). RNA quality was checked using a nanodrop, gel
   electrophoresis, and an Agilent Fragment Analyzer. RNA-seq libraries
   were generated by the Cornell TREx Facility using the NEBNext Ultra II
   Directional RNA Library Prep Kit (New England Biolabs) using 700ng
   input total RNA per sample. At least 20 M reads (2 × 150 nt PE) were
   generated on a NovaSeq (Illumina). Reads were trimmed to remove
   low-quality and adaptor sequences with TrimGalore (a wrapper for
   cutadapt and fastQC), requiring a minimum trimmed length of 50 nt.
   Reads that passed quality control were aligned to the reference genome
   (Ensembl 98.38) [[87]38] using STAR [[88]39], using ‘--quantMode
   GeneCounts’ to output counts per gene. SARTools [[89]40] and DESeq2
   [[90]41] were used to generate normalized counts and statistical
   analysis of differential gene expression. Genes with FDR control
   p-value ≤ 0.05 and absolute log2 fold change ≥ 0.5 were identified as
   differentially expressed genes (DEGs). All the genes detected in
   RNA-seq were listed in the Supplemental Dataset S2. Gene enrichment
   analysis using Hallmark and KEGG gene sets was performed using Gene Set
   Enrichment Analysis (GSEA) [[91]42]. The gene sets associated with
   inflammation were obtained from the GSEA hallmark database. The
   lysosome gene set is obtained from the GSEA KEGG database.

Brain proteomics analysis

   Mouse brain proteomics analysis was conducted as described previously
   [[92]43]. Briefly, mouse brain hemisphere tissue samples were
   lyophilized, weighed, homogenized in a methanol: water solvent mixture
   (5:2, v: v), and centrifuged to obtain a protein pellet. The protein
   pellet was dissolved in lysis buffer (8 M urea, 50 mM ammonium
   bicarbonate (AmBC), 150 mM NaCl), sonicated in ice-cold water, and
   clarified by centrifugation. Protein concentration was measured by a
   colorimetric DCA assay (Bio-Rad), and 40 µg of protein from each sample
   was used for protein digestion. Protein lysates were reduced,
   alkylated, diluted, and digested with trypsin/Lys-C (1:25, w: w)
   overnight, followed by quenching with TFA till pH < 2. Peptides were
   cleaned using a Waters HLB 96-well plate, dried, and reconstituted in
   2% ACN with 0.1% FA for LC-MS analysis on a Dionex Ultimate 3000
   RSLCnano system coupled with a Q Exactive HF-X Orbitrap MS. Peptides
   were separated using an Easy-Spray PepMap RSLC C18 column (2 μm, 100 Å,
   75 μm × 50 cm) with a 210-minute gradient at 55 °C and 0.25 µL/min flow
   rate. Data analysis was performed with Spectronaut software (v19.2,
   Biognosys) in DirectDIA mode, using the UniProt Mus musculus database
   alongside a custom contaminant library [[93]43]. Carbamidomethylation
   was set as the fixed modification. Methionine oxidation and N-terminal
   acetylation were included as variable modifications. A false discovery
   rate (FDR) cutoff of 0.01 was applied to peptide and protein
   identifications. Precursor intensities below 1,000 were excluded from
   downstream analysis. Identified proteins were listed in Supplemental
   Dataset S3. Statistical significance was determined using an unpaired
   Student’s t-test, with proteins exhibiting FDR < 0.05 and absolute
   log[2](FC) > 0.5 considered significantly altered. Pathway enrichment
   analysis was conducted via Metascape [[94]44].

Image analysis

   For the quantitative analysis of GFAP, CD68, and IBA1 levels in the
   brain sections, the fluorescence intensity was measured directly using
   ImageJ after a threshold application. To quantify the CD68 levels in
   microglia, IBA1-positive cells were selected, and CD68 signals within
   microglia were measured using the region of interest (ROI) tool in
   ImageJ. The lipofuscin quantification was determined by the red
   fluorescence signals (594 nm). For the quantitative analysis of CathD
   and LAMP1 levels in microglia and neurons, the IBA1 or NeuN-positive
   cells were selected, and CathD and LAMP1 within these cells were
   measured using the ImageJ ROI tool. Three brain sections per mouse,
   separated by 100 μm, were used for quantification. The mean from the
   three sections was used to be representative of each mouse. Data from
   ≥ 3 brains in each genotype were used for quantitative analysis and
   normalized to age-matched controls.

Statistical analysis

   Statistical analyses were performed using GraphPad Prism 10. All data
   are presented as mean ± SEM. Statistical significance was assessed by
   unpaired Student’s t-test (for two-group comparison) or one-way ANOVA
   tests with Bonferroni’s multiple comparisons (for multiple
   comparisons). P values less than or equal to 0.05 were considered
   statistically significant. *p < 0.05; **p < 0.01; ***p < 0.001.

Results

Loss of PGRN in the FVB/N strain background leads to earlier gliosis and
lysosomal abnormalities compared to the C57BL/6 background

   To explore the effect of strain background on PGRN-deficient
   phenotypes, we crossed Grn^−/− mice in the B6 background with wild-type
   (WT) FVB mice. Brain pathology was examined in Grn^−/− mice in the
   FVB/B6 hybrid background to determine the consequences of PGRN loss.
   Interestingly, we found a significant increase in glial activation,
   lipofuscin accumulation, and enlarged lysosomes in neurons and
   microglia in the 2-month-old Grn^−/− mice (Fig. [95]S1). To further
   confirm this phenotype, we backcrossed Grn^−/− mice into the FVB
   background for 6 additional generations and compared lipofuscin signals
   and glial activation in the 2-month-old Grn^−/− mice with littermate WT
   controls. Similar increases in glia activation and lipofuscin
   accumulation were observed (Fig. [96]1a and b), and most of the
   lipofuscin signals are found in NeuN + neurons and IBA1 + microglia but
   not in the GFAP + astrocytes (Fig. [97]S2). In contrast, PGRN loss only
   leads to a subtle increase in gliosis and lipofuscin accumulation in
   the 2-month-old B6 mice (Fig. [98]1a and b).

Fig. 1.

   [99]Fig. 1
   [100]Open in a new tab

   Glial activation and lipofuscin accumulation in 2-month-old
   PGRN-deficient FVB and B6 mice. (a) Immunostaining of GFAP, CD68, and
   IBA1 and auto-fluorescent lipofuscin signals in brain sections from
   2-month-old WT and Grn^−/− mice in the FVB or B6 strain backgrounds.
   Representative images from the cortex and the thalamus were shown for
   GFAP/CD68/IBA1 staining and lipofuscin signals. Scale bar = 100 μm
   (GFAP/CD68/IBA1) and scale bar = 10 μm (lipofuscin). (b) Quantification
   of GFAP, CD68, and IBA1 levels and auto-fluorescent lipofuscin signals
   for the experiment in (a). Data are presented as mean ± SEM from 3 mice
   per group (n = 3). 3 brain sections were analyzed for each mouse brain.
   *, p < 0.05, **, p < 0.01, ***, p < 0.001, ****, p < 0.0001, Student’s
   t-test

   To further compare lysosomal phenotypes caused by PGRN deficiency in
   FVB versus B6 mouse strains, we immunostained 2-month-old WT and
   Grn^−/− mice with antibodies against lysosomal protein Cathepsin D,
   together with microglial marker IBA1. A significant increase in
   Cathepsin D levels was detected in PGRN-deficient microglia and neurons
   (Fig. [101]2a and b). In addition, enlarged lysosomes are frequently
   found in PGRN-deficient microglia and neurons (Fig. [102]2a and b),
   indicating early lysosomal defects in both microglia and neurons upon
   PGRN loss in the FVB mice. In contrast, PGRN deficiency in the B6
   background only results in a subtle increase in Cathepsin D levels in
   neurons and microglia in the 2-month-old mice (Fig. [103]2a and b). In
   addition, because PGRN haploinsufficiency is a leading cause of FTLD,
   we examined the phenotype of the Grn^+/− mice in the FVB background. No
   obvious inflammation and lysosomal phenotypes were observed in these
   mice at 5 months old (Fig. [104]S3), indicating that partial loss of
   PGRN is not sufficient to induce inflammatory and lysosomal
   abnormalities at this age, even in the FVB background. Taken together,
   PGRN deficiency causes an earlier onset of inflammation and lysosomal
   abnormalities in the FVB background compared to the B6 background.

Fig. 2.

   [105]Fig. 2
   [106]Open in a new tab

   Lysosomal changes in 2-month-old PGRN-deficient FVB and B6 mice. (a, b)
   Immunostaining of Cathepsin D (CathD), NeuN, and IBA1 in brain sections
   from 2-month-old WT and Grn^−/− mice in FVB and B6 strain background.
   Representative images from the cortex and thalamus were shown. CathD
   signals in NeuN-positive neurons and IBA1-positive microglia were
   quantified (b). 3 sections per mouse brain and 3 mouse brains were used
   for each genotype. 15–20 microglia and 80–100 neurons per section were
   analyzed. Data are presented as mean ± SEM from 3 mice per group.
   n = 3, *, p < 0.05, **, p < 0.01, ***, p < 0.001. Student’s t-test.
   Scale bar = 10 μm

Loss of PGRN in the FVB/N mouse brain leads to exacerbated phenotypes during
aging and increased levels of phosphorylated TDP-43

   We further compared the PGRN deficiency phenotypes in 10-month-old FVB
   versus B6 mice. As expected, PGRN deficiency resulted in increased
   microglial and astrocyte activation as well as lipofuscin accumulation
   in both backgrounds (Figs. [107]3a and d and [108]4a and b). Apart from
   the comparable microglia activation in the cortex region between FVB
   and B6 Grn^−/− mice, changes in astrocyte activation and lipofuscin
   accumulation are more pronounced in both the thalamus and cortex of FVB
   Grn^−/− mice compared to B6 Grn^−/− mice, relative to WT (Figs. [109]3a
   and d and [110]4a and b). Additionally, levels of CathD in microglia
   and neurons in PGRN-deficient FVB and B6 mice were examined. CathD
   levels are significantly increased in the PGRN knockout mouse brain
   compared to WT in both backgrounds, with a further increase in FVB mice
   compared to the B6 mice in the thalamus but not in the cortex
   (Fig. [111]4c and d and Fig.[112]S4a, [113]4b). Furthermore, we
   analyzed the levels of GPNMB and Galectin-3, two markers associated
   with microglial activation and lysosomal damage, which are known to be
   upregulated in PGRN-deficient microglia in mouse brains and FTLD-GRN
   brains [[114]44–[115]48]. These markers are increased in microglia in
   the thalamus of Grn^−/− mice in both backgrounds, with much more
   significant increases observed in FVB mice (Fig. [116]4c and f). These
   results suggest that the FVB background exacerbates gliosis and
   lysosomal phenotypes observed under PGRN-deficient conditions,
   especially in the thalamus region.

Fig. 3.

   [117]Fig. 3
   [118]Open in a new tab

   Glial activation in 10-month-old PGRN-deficient FVB and B6 mice. (a)
   Immunostaining of GFAP, CD68, and IBA1 in brain sections from
   10-month-old WT and Grn^−/− mice in the FVB or B6 strain backgrounds.
   Representative images from the cortex and the thalamus were shown for
   GFAP/CD68/IBA1 staining (a). Scale bar = 100 μm. (b) Quantification of
   GFAP, CD68, and IBA1 levels in (a). Data are presented as mean SEM from
   3–4 mice per group (n = 3–4). 3 brain sections were analyzed for each
   mouse brain. *, p < 0.05, **, p < 0.01, ***, p < 0.001, ****,
   p < 0.0001, Student’s t-test. (c, d) Western blot analysis of GFAP in
   the thalamus and cortex of WT and Grn^−/− mice in FVB and B6
   backgrounds. The GFAP levels were quantified by using ImageJ and were
   normalized to the levels of the loading control GAPDH (D). Data are
   presented as mean ± SEM from 3–4 mice per group (n = 3–4). *, p < 0.05,
   ***, p < 0.001, ****, p < 0.0001. Student’s t-test

Fig. 4.

   [119]Fig. 4
   [120]Open in a new tab

   Lysosomal changes in 10-month-old PGRN-deficient FVB and B6 mice. (a,
   b) Lipofuscin accumulation in 10-month-old WT and Grn^−/− mice in FVB
   and B6 backgrounds. Representative images from the cortex and the
   thalamus were shown. Scale bar = 10 μm. Auto-fluorescent lipofuscin
   signals were quantified (B). (c, d) Western blot analysis of CathD and
   Galectin3 in the thalamus or cortex of WT and Grn^−/− mice in FVB and
   B6 backgrounds. The levels of both precursor (Pro-CathD) and mature
   form (Mat-CathD) of CathD, as well as Galectin3, were quantified by
   using ImageJ and were normalized to GAPDH levels (d). Data are
   presented as mean ± SEM from 3–4 mice per group (n = 3–4). **,
   p < 0.01, ***, p < 0.001, ****, p < 0.0001. Student’s t-test. (e, f)
   Immunostaining of Galectin-3, GPNMB, and IBA1 in brain sections from
   10-month-old WT and Grn^−/− mice in the FVB and the B6 strain
   background. Representative images from the thalamus were shown.
   Galectin-3 and GPNMB levels in microglia in the thalamus were
   quantified in (f). Data are presented as mean ± SEM from 3–4 mice per
   group (n = 3–4), *, p < 0.05, **, p < 0.01, ***, p < 0.001. Student’s
   t-test. Scale bar = 10 μm

   TDP-43 aggregation is a hallmark of FTLD-GRN. We asked whether PGRN
   loss triggers TDP-43 pathology in FVB mice. Total and phosphorylation
   levels of TDP-43 in RIPA-soluble and urea-soluble fractions from 2- and
   10-month-old Grn^−/− mice in both FVB and B6 backgrounds were analyzed.
   At 2 months old, PGRN deficiency barely affected the TDP-43 levels,
   solubility, and phosphorylation in both backgrounds (Fig. [121]5a and
   b). Interestingly, PGRN loss in the FVB background led to a significant
   increase in the levels of phosphorylated TDP-43 in urea-soluble
   fractions of 10-month-old mice (Fig. [122]5c and d), whereas PGRN loss
   in the B6 background had little effect (Fig. [123]5c and d), indicating
   that PGRN deficiency in the FVB background leads to enhanced TDP-43
   pathology.

Fig. 5.

   [124]Fig. 5
   [125]Open in a new tab

   PGRN-deficiency in the FVB strain leads to enhanced TDP-43 pathology.
   (a-d) Analysis of TDP-43 solubility and phosphorylation in cortical
   lysates from 2- and 10-month-old WT and Grn^−/− mice in FVB and B6
   strain backgrounds. Total TDP-43 levels and phosphorylated TDP-43
   (pS409/410) levels in the RIPA-soluble and urea-soluble fractions of 2-
   and 10-month-old mice were quantified in (b) and (d), respectively.
   Data are presented as mean ± SEM from 3–4 mice per group (n = 3–4). *,
   p < 0.05, **, p < 0.01. Student’s t-test

PGRN physically interacts with sPLA2-IIA and regulates its activity in vitro

   To understand the mechanisms of PGRN actions and dissect pathways
   affecting PGRN deficient phenotypes, we searched for PGRN interactors
   using a SILAC (stable isotope labeling by amino acids in cell culture)
   based proteomic screen (Fig. [126]6a). In addition to previously
   identified PGRN interactors, prosaposin (PSAP) [[127]49] and folding
   chaperones (BIP, PDIA3, DJC10) [[128]50], sPLA2-IIA (secreted
   phospholipase A2 type IIA, encoded by the Pla2g2a gene) was identified
   as one of the top hits from the screen (Fig. [129]6a, Supplementary
   Dataset 1). sPLA2-IIA is a secreted phospholipase of the sPLA2 family
   that catalyzes the hydrolysis of the sn-2 position of
   glycerophospholipids to generate fatty acids and lysophospholipids,
   many of which play a critical role in inflammation [[130]51].
   Interestingly, the B6 mouse strain harbors a naturally occurring
   mutation in Pla2g2a, which leads to the loss of the sPLA2-IIA protein
   in this background [[131]52], while FVB mice express wild-type
   sPLA2-IIA. We hypothesized that sPLA2-IIA might explain the strain
   background difference observed for PGRN deficiency. To explore the
   relationship between PGRN and sPLA2-IIA, the PGRN-sPLA2-IIA interaction
   was first verified by co-immunoprecipitation (IP) with overexpressed
   human PGRN and sPLA2-IIA in both the medium and cell lysates
   (Fig. [132]6b). This interaction is conserved with mouse PGRN and mouse
   sPLA2-IIA (Fig. [133]6c). Next, we determined the functional
   significance of PGRN interaction with sPLA2-IIA. To test whether PGRN
   could directly regulate sPLA2-IIA activities, we purified recombinant
   sPLA2-IIA from the medium of transfected HEK293T cells and performed in
   vitro assays with colorimetric substrates of sPLA2 in the absence or
   presence of recombinant PGRN. We found that the presence of PGRN
   significantly reduced the activity of sPLA2-IIA at 1:1 and 1:2 molar
   ratios compared to the control reaction with sPLA2-IIA alone or with
   sPLA2-IIA and bovine serum albumin (BSA) at a 1:1 ratio (Fig. [134]6d).

Fig. 6.

   [135]Fig. 6
   [136]Open in a new tab

   PGRN physically interacts with sPLA2-IIA and regulates sPLA2-IIA levels
   and activity. (a) Schematic illustration of the SILAC experiment
   searching for PGRN interactors in HEPG2 cells. Proteins with more than
   4 peptides and p-value < 0.01 and Log2 (GFP/GFP-PGRN) ratio <-1 were
   labeled in green. (b) Human PGRN and sPLA2-IIA interact when
   overexpressed in HEK293T cells. Human sPLA2-IIA-myc and human
   PGRN-expressing constructs were transfected into HEK293T cells as
   indicated. Lysates and conditioned medium were prepared 3 days later
   and immunoprecipitated with anti-Myc antibodies. (c)
   Mouse-sPLA2-IIA-myc and FLAG-mouse-PGRN expressing constructs were
   transfected into HEK293T cells as indicated. Lysates and conditioned
   medium were prepared 3 days later and immunoprecipitated with anti-Myc
   antibodies. (d) PGRN inhibits sPLA2-IIA activities in vitro.
   Recombinant sPLA2-IIA protein was incubated with substrates in the
   absence or presence of recombinant PGRN at a 1:1 or 1:2 molar ratio,
   and sPLA2-IIA activities were measured. Data are presented as
   mean ± SEM from three independent experiments (n = 3). *, p < 0.05; **,
   p < 0.01, Student’s t-test. (e) The levels of sPLA2-IIA in tissue
   lysates from 2-month-old FVB mice were analyzed by western blot. (f)
   The levels of sPLA2-IIA are upregulated in the duodenum in
   PGRN-deficient mice. The duodenum of 2-month-old WT B6 mice and WT
   (Pla2g2a^+/+Grn^+/+) and Grn^−/− (Pla2g2a^+/+Grn^−/−) mice in the FVB
   background were lysed, and the same amount of the total protein was
   used for western blot to measure sPLA2-IIA levels. sPLA2-IIA levels in
   the mouse duodenum were quantified. Data are presented as mean ± SEM
   from 5 mice per group (n = 5). *, p < 0.05, Student’s t-test. (g, h)
   Serum levels of prostaglandin and C18:1, C18:2, and C18:3 fatty acids
   are upregulated in PGRN-deficient mice. Prostaglandin (predominantly
   PGE2 and PGD2, which have the same mass) concentration in serum and
   cortical lysates was measured by mass spectrometry using PGE2 as the
   standard. The levels of C18:1, C18:2, and C18:3 fatty acids were
   measured by mass spectrometry, and the relative abundance in WT and
   Grn^−/− mice was shown. Data are presented as mean ± SEM from six mice
   per group (n = 6). *, p < 0.05, **, p < 0.01, ***, p < 0.001, and ****,
   p < 0.0001, Student’s t-test

PGRN loss results in increased levels of sPLA2-IIA in the gut and elevated
levels of prostaglandins and fatty acids in the serum

   We next examined the effect of PGRN loss on sPLA2-IIA levels and
   functions in FVB mice. Consistent with previously published data
   [[137]53], we found that the expression levels of sPLA2-IIA are not
   detectable in the brain but high in the intestine (Fig. [138]6e).
   Recently, sPLA[2]-IIA has been shown to regulate the gut microbiome,
   inflammation, and metabolism in the intestine [[139]54–[140]56].
   Interestingly, we found that PGRN loss leads to a significant increase
   in sPLA2-IIA protein levels in the duodenum (Fig. [141]6f). Since
   sPLA2-IIA is a secreted enzyme, we next attempted to determine changes
   in the levels of secreted sPLA2-IIA. Unfortunately, we could not
   measure serum levels of sPLA2-IIA because we failed to identify a
   commercial ELISA kit specific to mouse sPLA2-IIA (data not shown).
   Thus, we instead decided to measure alterations in the levels of
   downstream products of sPLA2, prostaglandins, and free fatty acids
   [[142]51, [143]57, [144]58]. We found a significant increase in the
   serum levels of prostaglandins PGE2 and PGD2, as well as C18:1, C18:2,
   and C18:3 fatty acids in PGRN-deficient mice (Fig. [145]6g).
   Interestingly, this increase is not found in the cortex lysates from
   PGRN-deficient mice (Fig. [146]6h).

sPLA2 inhibitors ameliorate phenotypes associated with PGRN loss in the mouse
brain

   Because sPLA2-IIA is an important mediator of inflammation through the
   generation of arachidonic acid (AA) and other fatty acid signaling
   molecules [[147]51, [148]57, [149]58], we hypothesized that PGRN might
   regulate inflammation through sPLA2-IIA. To determine whether increased
   sPLA2-IIA activities contribute to the inflammation and lysosomal
   phenotypes in PGRN-deficient mice, we treated 5-week-old
   Pla2g2a^+/+Grn^−/− mice in the mixed FVB/B6 background with
   [150]LY333013, a broad inhibitor of the sPLA2 family phospholipases, or
   KH064, a relatively specific inhibitor of sPLA2-IIA [[151]57]. Both
   inhibitors significantly decreased sPLA2 activities toward a commercial
   substrate in the serum and in the cortical lysates, indicating that the
   inhibitors have crossed the blood-brain barrier (Fig. [152]7a).
   Strikingly, the inhibition of sPLA2 not only rescues the microgliosis
   and astrogliosis phenotypes but also lipofuscin accumulation in Grn^−/−
   mice in the Pla2g2a^+/+ background at 2-month-old (Fig. [153]7b and c).
   Enlarged lysosomes and increased levels of cathepsin D in neurons and
   microglia of Pla2g2a^+/+Grn^−/− mice are also rescued by treatment with
   sPLA2s inhibitors (Fig. [154]7d and e). These results support that the
   increase in sPLA2-IIA activity might contribute to enhanced
   inflammation and lysosomal abnormalities in the Pla2g2a^+/+Grn^−/−
   mice.

Fig. 7.

   [155]Fig. 7
   [156]Open in a new tab

   sPLA2 inhibitor treatment rescues PGRN deficiency phenotypes in mice.
   5-week-old Pla2g2a^+/+Grn^+/+ and Pla2g2a^+/+Grn^−/− mice in FVB/B6
   mixed background were treated with corn oil control or sPLA2
   inhibitors, [157]LY333013 or KH064 for 3 weeks. (a) Measurement of
   total sPLA2 activity in the serum and cortical lysates from untreated
   and [158]LY333013 or KH064-treated mice. Data are presented as
   mean ± SEM from five mice per group (n = 5). *, p < 0.05; **, p < 0.01,
   Student’s t-test. (b) Brain sections were immunostained with anti-GFAP,
   CD68, and IBA1 antibodies, and representative images from the cortex
   region were shown (Scale bar = 100 μm). Another set of brain sections
   was stained with Hoechst and auto-fluorescent lipofuscin signals (red)
   in the thalamus region were imaged (scale bar = 10 μm). (c)
   Quantification of GFAP, CD68, and IBA1 levels and auto-fluorescent
   lipofuscin signals for the experiment in (b). Data are presented as
   mean ± SEM from five mice per group (n = 5). 3 sections were analyzed
   for each mouse brain. *, p < 0.05; **, p < 0.01, Student’s t-test. (d)
   Brain sections were immunostained with anti-CathD, NeuN, and IBA1
   antibodies, and representative images from the cortex region were shown
   (Scale bar = 10 μm). (e) CathD signals in IBA1-positive microglia and
   NeuN-positive neurons were quantified. 10–20 confocal images were
   randomly captured from each brain section. Ten to twenty microglia and
   fifty to sixty neurons were quantified per section. 3 sections were
   analyzed for each mouse brain. Data are presented as mean ± SEM from
   3–4 mice per group (n = 3–4). *, p < 0.05; **, p < 0.01, Student’s
   t-test

AAV-mediated sPLA2-IIA overexpression leads to enhanced glial activation and
lysosomal abnormalities in PGRN-deficient mice

   Next, we test whether sPLA2-IIA expression is sufficient to drive
   inflammation and lysosomal abnormalities. We found that expression of
   either mouse or human sPLA2-IIA in WT B6 mice does not cause glial
   activation or lipofuscin accumulation (Fig.[159]S5 and Fig.[160]S6). We
   also generated Pla2g2a^+/+Grn^−/− mice in the B6 background by crossing
   FVB Grn^−/− mice (Pla2g2a^+/+Grn^−/−) with B6 Grn^−/− mice
   (Pla2g2a^−/−Grn^−/−) and compared their phenotypes with those of
   Pla2g2a^−/−Grn^−/− mice in the B6 background. We found that expression
   of sPLA2-IIA at endogenous levels does not enhance the glial activation
   and lipofuscin accumulation in 2-month-old Grn^−/− mice in the B6
   background (Fig. [161]8a and b), indicating that basal levels of
   sPLA2-IIA are not sufficient to affect PGRN deficiency phenotypes. To
   further investigate the role of sPLA2-IIA in driving PGRN deficiency
   phenotypes, we overexpressed sPLA2-IIA in WT and Grn^−/− mice in the B6
   strain background via AAV injection. Postnatal (P0) pups were injected
   with AAV9 viruses overexpressing sPLA2-IIA or buffer control orbitally.
   Mice were collected at 2 months old, and gliosis and lysosomal
   phenotypes were examined in brain sections. We found that sPLA2-IIA is
   highly expressed in the neurons in the cortex, but its expression in
   the thalamus region was minimal (Fig. [162]S7a). An obvious
   colocalization between sPLA2-IIA and PGRN was observed (Fig. [163]S7b),
   indicating that sPLA2-IIA is localized in the lysosomal compartment in
   neurons. Interestingly, we found that sPLA2-IIA overexpression in the
   cortex led to enhanced gliosis and lipofuscin accumulation in Grn^−/−
   mice but not in WT mice (Figs. [164]8c and d and [165]9a and b),
   suggesting that elevated sPLA2-IIA activity in the brain can only drive
   gliosis or lysosomal phenotypes when combined with PGRN loss. More
   intriguingly, the increased lipofuscin accumulation was found
   specifically in the microglia but not in neurons from the Grn^−/− mouse
   brain cortex region (Fig. [166]9a and b). Consistently, the levels of
   lysosomal enzyme CathD were significantly increased in microglia in
   PGRN-deficient mice upon sPLA2-IIA overexpression (Fig. [167]9c and d).
   These results further demonstrate that elevated sPLA2-IIA activities in
   PGRN-deficient mice enhance inflammation and lysosomal abnormalities.

Fig. 8.

   [168]Fig. 8
   [169]Open in a new tab

   AAV9-mediated overexpression of sPLA2-IIA leads to enhanced glial
   activation in 2-month-old B6 Grn^−/− mice. (a, b) Expression of
   sPLA2-IIA at endogenous levels in Grn^−/− B6 mice does not affect glial
   activation and lipofuscin accumulation. Immunostaining of GFAP, CD68,
   and IBA1 and autofluorescent lipofuscin signals in brain sections from
   2-month-old Pla2g2a^−/−Grn^−/− and Pla2g2a^+/+Grn^−/− littermate mice
   in the B6 background. Representative images from the cortex and the
   thalamus were shown for GFAP/CD68/IBA1 staining and lipofuscin signals.
   Scale bar = 100 μm (for GFAP/CD68/IBA1) and 10 μm (for lipofuscin).
   GFAP, CD68, and IBA1 levels and auto-fluorescent lipofuscin signals
   were quantified in (b). Data are presented as mean ± SEM from 4 mice
   per group (n = 4). 3 brain sections were analyzed for each mouse brain.
   (c, d) AAV9-mediated overexpression of sPLA2-IIA leads to enhanced
   glial activation in 2-month-old Grn^−/− mice in B6 mice. B6 WT
   (Pla2g2^−/−Grn^+/+) and Grn^−/− (Pla2g2a^−/−Grn^−/−) pups at postnatal
   day (P0) were injected with AAV9-sPLA2-IIA or the buffer control, and
   the mice were collected at 2 months old. (c) Brain sections were
   immunostained with anti-GFAP, CD68, and IBA1 antibodies, and
   representative images from the cortex and thalamus regions were shown
   (Scale bar = 100 μm). (d) Quantification of GFAP, CD68, and IBA1 levels
   for the experiment in (c). Data are presented as mean ± SEM from three
   to five mice per group (n = 3–5). Three sections were analyzed for each
   mouse brain. *, p < 0.05; **, p < 0.01; ****, p < 0.0001, Student’s
   t-test

Fig. 9.

   [170]Fig. 9
   [171]Open in a new tab

   AAV9-mediated overexpression of sPLA2-IIA leads to lipofuscin
   accumulation and lysosomal enlargement in microglia in 2-month-old B6
   Grn^−/− mice. B6 WT (Pla2g2^−/−Grn^+/+) and Grn^−/−
   (Pla2g2a^−/−Grn^−/−) pups at postnatal day (P0) were injected with
   AAV9-sPLA2-IIA or the buffer control, and the mice were collected at 2
   months old. (a) Brain sections were stained with NeuN and IBA1, and
   auto-fluorescent lipofuscin signals (blue) in the cortex region were
   imaged (scale bar = 10 μm). (b) Quantification of autofluorescent
   lipofuscin signals in the cortex. Data are presented as mean ± SEM from
   three to five mice per group (n = 3–5). Three sections were analyzed
   for each mouse brain. *, p < 0.05; **, p < 0.01; ****, p < 0.0001,
   Student’s t-test. (c) Brain sections were immunostained with
   anti-CathD, NeuN, and IBA1 antibodies, and representative images from
   the cortex region were shown (Scale bar = 10 μm). (d) CathD signals in
   IBA1-positive microglia and NeuN-positive neurons were quantified.
   10–20 confocal images were randomly captured from each brain section.
   Fifteen to thirty microglia and fifty to sixty neurons were quantified
   per Sect. 3 sections were analyzed for each mouse brain. Data are
   presented as mean ± SEM from 3–5 mice per group (n = 3–5). *, p < 0.05;
   **, p < 0.01; ****, p < 0.0001, Student’s t-test

PGRN deficiency in FVB/N and C57BL/6 mouse strains leads to distinct
alterations in transcriptome and proteome

   Since endogenous sPLA2-IIA expression in the B6 background is not
   sufficient to cause PGRN deficiency phenotypes, we performed an RNA-seq
   analysis with RNAs isolated from cortical regions of 5-month-old WT and
   Grn^−/− mice in a B6 versus FVB/B6 mixed background to gain more
   insight into the mechanisms driving the phenotypic differences caused
   by PGRN loss in FVB and B6 backgrounds. These analyses revealed
   substantial differences in gene expression between the two strain
   backgrounds (Fig. [172]10a and Supplementary Dataset 2). Additionally,
   only 10 genes overlap in the lists of differentially expressed genes
   (DEGs) identified in Grn^−/− mice compared to WT in the two strain
   backgrounds (Fig. [173]10b and c). These genes upregulated in Grn^−/−
   mice are all related to immune functions (Fig. [174]10c). Other genes
   that are specifically impacted by PGRN deficiency in each strain
   background are depicted in the heatmaps (Fig. [175]10d and g and
   Supplementary Dataset 2). Pathway analysis further revealed that
   different pathways are enriched in PGRN-deficient mice across two
   strain backgrounds (Fig. [176]10h). Nevertheless, lysosome and
   inflammation pathways are among the most upregulated pathways in both
   strain backgrounds upon PGRN loss (Fig. [177]10h), consistent with the
   known roles of PGRN. Notably, the oxidative phosphorylation pathway is
   significantly downregulated under PGRN deficiency in the FVB/B6 mixed
   background, whereas it is upregulated in the B6 background
   (Fig. [178]10h and i), suggesting that it might contribute to the
   phenotypic differences caused by PGRN loss in these two strain
   backgrounds.

Fig. 10.

   [179]Fig. 10
   [180]Open in a new tab

   Gene expression changes caused by PGRN deficiency in the FVB/B6 and B6
   mouse strains. Total RNAs were extracted from the cortex of 5-month-old
   WT and Grn^−/− mice in B6 and FVB/B6 mixed background, and the RNA-seq
   was performed to analyze gene expression changes (n = 3). (a, b) Venn
   diagram showing the overlap of the differentially expressed genes
   (DEGs) (FDR ≤ 0.05,|Log2FC| ≥ 0.5) identified in the indicated
   comparisons. (c) Heatmap showing the expression changes of genes
   significantly upregulated in Grn^−/− mice compared to WT in both B6 and
   FVB/B6 mixed backgrounds. DEGs with FDR ≤ 0.05 and log2(FC) ≥ 0.5 were
   plotted. (d) Heatmap showing the expression changes of genes
   specifically upregulated in Grn^−/− mice compared to WT in the FVB/B6
   mixed background. DEGs with FDR ≤ 0.05 and log2(FC) ≥ 1.0 were plotted.
   (e) Heatmap showing the expression changes of genes specifically
   downregulated in Grn^−/− mice compared to WT in FVB/B6, the mixed
   background. DEGs with FDR ≤ 0.05 and log2(FC) ≤ -0.5 were plotted. (f)
   Heatmap showing the expression changes of genes specifically
   upregulated in Grn^−/− mice compared to WT in the B6 background. DEGs
   with FDR ≤ 0.05 and log2(FC) ≥ 0.5 were plotted. (g) Heatmap showing
   the expression changes of genes specifically downregulated in Grn^−/−
   mice compared to WT in the B6 background. DEGs with FDR ≤ 0.05 and
   log2(FC) ≤ -1.0 were plotted. (h) GSEA pathway enrichment analysis to
   identify the significantly upregulated and downregulated gene sets (FDR
   q-value < 0.1) in the indicated comparisons. The gene sets with
   normalized enrichment score (NES) >1.0 and those with NES < -1 are
   listed as upregulated or downregulated in the indicated groups. The
   gene sets associated with inflammation are highlighted in orange, and
   the lysosome gene set is highlighted in blue. (i) Heatmap showing the
   expression changes of the genes involved in oxidative phosphorylation

   We also performed proteomic analyses to identify protein changes
   triggered by PGRN loss in FVB and B6 backgrounds. Proteomic comparison
   of proteins in the brain lysates of 2-month-old FVB versus B6 mice
   identified approximately 1,600 DEGs (Fig. [181]11a and Supplementary
   Dataset 3). PGRN loss leads to more protein changes in the FVB mice
   compared to B6, as expected (Fig. [182]11b and Supplementary Dataset
   3). However, minimal overlap between protein changes is observed in the
   two strain backgrounds upon PGRN loss (Fig. [183]11b). Seventeen
   proteins are significantly increased in Grn^−/− mice in both
   backgrounds (Fig. [184]11b and c), such as solute carrier family
   members SLC39A7 and SLC7A14, which function as zinc and amino acid
   transporters, respectively. Three proteins involved in lipid
   metabolism, including palmitoyl-protein thioesterase 1 (PPT1),
   sphingomyelin phosphodiesterase 1 (SMPD1), and N-acylsphingosine
   amidohydrolase 1 (ASAH1), were decreased in both FVB and B6 Grn^−/−
   mice compared to WT (Fig. [185]11f). Other proteins specifically
   affected by PGRN deficiency in each strain background are displayed in
   the heatmaps (Fig. [186]11d and e). Furthermore, pathway analysis
   revealed that inorganic ion transmembrane transport is the only pathway
   upregulated by PGRN loss in both backgrounds (Fig. [187]11g).
   Interestingly, multiple lipid metabolism pathways are significantly
   upregulated in Grn^−/− mice in the FVB but not B6 background, while
   pathways associated with mitochondrial function, such as aerobic
   respiration and respiratory electron transport, ATP metabolic process,
   and regulation of reactive oxygen species metabolic process, were
   specifically upregulated by PGRN deficiency in B6 but not FVB
   background (Fig. [188]11g and h). These findings suggest that
   alterations in lipid metabolism and mitochondrial function might be
   associated with the phenotypic differences caused by PGRN loss in the
   two strain backgrounds.

Fig. 11.

   [189]Fig. 11
   [190]Open in a new tab

   Proteomic changes caused by PGRN deficiency in the FVB and B6 mouse
   strains. Total proteins were extracted from brain hemispheres of
   2-month-old WT and Grn^−/− mice in FVB and B6 backgrounds, and the
   proteomic analysis was performed to analyze protein changes (n = 4).
   (a, b) Venn diagram showing the overlap of the significantly changed
   proteins (FDR ≤ 0.05,|log2FC| ≥0.5) identified in the indicated
   comparisons. (c) Heatmap showing the expression of the proteins that
   were significantly upregulated (FDR ≤ 0.05, log2FC ≥ 0.5) in Grn^−/−
   mice compared to WT in both FVB and B6 backgrounds. (d) Heatmap showing
   the expression of the proteins (FDR ≤ 0.05, log2FC ≥ 1) specifically
   upregulated in FVB Grn^−/− compared with WT. (e) Heatmap showing the
   expression of the proteins (FDR ≤ 0.05, log2FC ≥ 1) specifically
   upregulated in B6 Grn^−/− mice compared with WT. (f) Heatmap showing
   the expression of the significantly downregulated proteins identified
   in FVB Grn^−/− mice compared with WT (top) and the downregulated
   proteins identified in B6 Grn^−/− mice compared with WT (bottom)
   (FDR ≤ 0.05 and log2FC ≤ -0.5). The proteins downregulated in Grn^−/−
   mice in both backgrounds are highlighted in red. (g) Upregulated
   pathways in Grn^−/− mice compared to WT mice in FVB (top) or B6
   (bottom) background. Pathway enrichment analysis using gene ontology
   biology process (GOBP) and Reactome gene sets was performed using
   Metascape. The enriched pathways with p-value <0.01 were shown. The
   pathways associated with lipid metabolism, mitochondrial function, and
   inorganic ion transmembrane transport are highlighted in green, orange,
   and blue, respectively. (h) Heatmap showing the changes in the proteins
   involved in aerobic respiration and the aerobic respiratory electron
   pathway. Only those proteins with an intensity above 1,000 are
   displayed in the heatmap

   Taken together, the above results suggest that PGRN loss causes
   substantial differences in alterations in transcriptome and proteome in
   the FVB and B6 strain background, which could potentially modify PGRN
   deficiency phenotypes.

Discussion

sPLA2 activities modify PGRN-deficient phenotypes in the mouse brain

   sPLA2-IIA is a member of a secreted PLA2 protein family (sPLA2), which
   catalyzes the hydrolysis of the sn-2 position of glycerophospholipids.
   It is an important mediator of inflammation through the generation of
   arachidonic acids (AAs) and other fatty acid signaling molecules
   [[191]51, [192]57, [193]58]. The sPLA2 family consists of 11 members,
   each of which has distinct tissue-specific expression patterns and
   slight differences in substrate specificity [[194]51, [195]57,
   [196]58]. In this study, we have identified sPLA2-IIA as a novel
   interactor for PGRN and shown that PGRN inhibits sPLA2-IIA activity in
   vitro (Fig. [197]6a and d). In addition, PGRN loss results in increased
   levels of sPLA2-IIA in the gut and elevated levels of prostaglandins
   and fatty acids in the serum (Fig. [198]6e and h). Among sPLA2 members,
   only sPLA2-IIA is known to be differentially expressed between FVB and
   B6, with a naturally occurring mutation and subsequent loss of
   sPLA2-IIA protein in the B6 background. We found that PGRN deficiency
   in the sPLA2-IIA expressing FVB mouse strain shows a much earlier onset
   and more severe phenotypes compared to the B6 background
   (Fig. [199]1-[200]5). Additionally, AAV-mediated sPLA2-IIA
   overexpression in the cortex leads to enhanced phenotypes in
   PGRN-deficient B6 mice, indicating that increased sPLA2 activities in
   the brain exacerbate the phenotypes associated with PGRN loss. These
   findings led us to hypothesize that elevated sPLA2-IIA activity might
   contribute to the enhanced phenotype in the FVB strain with PGRN
   deficiency.

   Elevated sPLA2 levels have been implicated in multiple inflammatory
   disorders [[201]59]. The two sPLA2 inhibitors KH064 and [202]LY333013
   have been developed to treat sPLA2-associated diseases [[203]59].
   Remarkably, KH064 or [204]LY333013 treatment significantly alleviates
   gliosis and lysosomal abnormalities in PGRN-deficient mouse brains in
   the FVB background (Fig. [205]7). While KH064 has been reported as a
   specific inhibitor of sPLA2-IIA [[206]60], [207]LY333013 is a broad
   inhibitor that targets several sPLA2 members, including sPLA2-V and
   sPLA2-X [[208]61]. At this stage, we can not rule out the possibility
   that these two inhibitors might also target other sPLA2 family members
   or even other unrelated proteins to regulate inflammation and lysosomal
   activities in PGRN-deficient mice, since sPLA2-IIA level is low in the
   brain. Nevertheless, our results suggest that the two sPLA2 inhibitors
   might be a valid therapeutic approach for FTLD-GRN, and possibly also
   other neurodegenerative diseases, to restrict neuroinflammation
   [[209]62]. It will be worth testing the effect of sPLA2 inhibitors in
   other mouse models of neuroinflammation and neurodegeneration in the
   future.

   In humans, although polymorphisms in the Pla2g2a gene affecting
   sPLA2-IIA levels have been associated with metabolic disorders, such as
   coronary heart disease and type II diabetes [[210]63–[211]65], no
   genetic link between Pla2g2a and FTLD or other neurodegenerative
   diseases has been reported. In the mouse brain, we have failed to
   detect endogenous sPLA2-IIA signals by western blot (Fig. [212]6e) or
   immunostaining (data not shown). However, sPLA2-IIA protein was
   detected in the brain lysate when human sPLA2-IIA is expressed under
   its native promoter in the Tg-hPla2g2a mice (Fig. [213]S6b). Pevious
   studies have shown that sPLA2-IIA is expressed by astrocytes [[214]66,
   [215]67] and endothelial cells [[216]68]. Secreted sPLA2-IIA has been
   shown to cause neuronal toxicity [[217]67], stimulate phagocytosis,
   proliferation, and inflammatory responses of microglia-derived BV2
   cells [[218]69], and proliferation of an astrocytoma cell line
   [[219]70]. sPLA2-IIA levels are increased after spinal cord injury, and
   sPLA2-IIA has a detrimental effect [[220]71]. In addition, sPLA2-IIA
   mRNA is up-regulated in Alzheimer’s disease (AD) brains, and
   sPLA2-IIA-immunoreactive astrocytes are associated with amyloid β
   (Aβ)-containing plaques in AD [[221]62]. Oligomeric Aβ1–42 and
   interleukin-1β (IL-1β) were shown to induce sPLA2-IIA mRNA expression
   [[222]62]. Thus, although sPLA2-IIA levels are low in the brain,
   brain-expressed sPLA2-IIA might still be subject to regulation by PGRN
   to regulate brain inflammatory phenotypes, especially in humans. On the
   other hand, the highest level of sPLA2-IIA expression is found in the
   gut (Fig. [223]6e), and sPLA[2]-IIA can regulate gut microbiota,
   inflammation, and metabolism [[224]54–[225]56]. It also functions as an
   intestinal stem cell niche factor to regulate intestine homeostasis,
   inflammation, and cancer [[226]72]. An increase in sPLA2-IIA levels in
   the duodenum was found under PGRN-deficient conditions (Fig. [227]6f).
   Alterations in the gut can affect brain pathology through many
   different mechanisms [[228]73, [229]74]. Therefore, another possibility
   is that PGRN may regulate brain inflammation phenotypes by modulating
   sPLA2-IIA function and levels in the gut. Additionally, an increase in
   the levels of downstream products of sPLA2-IIA, prostaglandin, and free
   fatty acids is observed in the serum of PGRN-deficient mice
   (Fig. [230]6G), which could pass blood-brain barrier to affect brain
   phenotypes. Future studies are needed to pinpoint the actions of PGRN
   and sPLA2-IIA in the periphery versus in the brain.

Connections between inflammation and lysosomal abnormalities

   Intriguingly, sPLA2-IIA overexpression not only leads to glial
   inflammation but also lysosomal abnormalities and lipofuscin
   accumulation and sPLA2-IIA inhibitors rescue both lipofuscin
   accumulation and lysosomal abnormalities under PGRN deficient
   conditions (Figs. [231]7 and [232]8, and Fig. [233]9). Though the
   mechanisms by which sPLA2-IIA regulates lysosomal phenotypes remain to
   be investigated, many recent studies have established a close link
   between lysosomal dysfunction and inflammation. For example, lysosomal
   membrane permeability (LMP) often results in the release of lysosomal
   enzymes and substrates into the cytosol, which often leads to
   inflammasome activation [[234]75–[235]77]. In addition, the activation
   of TFE3/TFEB, the main transcriptional factor involved in lysosome
   biogenesis, drives the expression of not only lysosomal genes but also
   inflammation genes in several studies [[236]78, [237]79]. Furthermore,
   substrate accumulation and lysosome stress can activate Stat3, which
   mediates the transcription of many genes involved in inflammatory
   responses [[238]80]. However, whereas sPLA2-IIA is important for
   inflammatory responses by producing lipid mediators of inflammation,
   its role in the lysosomal compartment is completely unexplored.
   Interestingly, AAV9-mediated sPLA2-IIA overexpression in neurons in
   PGRN-deficient mouse brains triggers lipofuscin accumulation and
   lysosomal enlargement specifically in microglia (Fig. [239]9). Since
   sPLA2-IIA is a secreted enzyme, hydrolysis of lipids in extracellular
   vesicles or the extracellular leaflet of the plasma membrane may lead
   to the production of AAs and other inflammatory lipid mediators, such
   as prostaglandins [[240]51, [241]57, [242]58]. The free fatty acids or
   other molecules produced by sPLA2-IIA could be uptaken by microglia to
   cause lysosomal abnormalities under PGRN-deficient conditions.
   sPLA2-IIA has also been shown to have catalytically independent
   functions, such as signaling through its receptors [[243]51, [244]57,
   [245]58]. Thus, sPLA2-IIA might also affect microglia lysosomal
   activities through its signaling functions or other mechanisms.

Mouse strain background modifies PGRN deficiency phenotypes

   Previous studies have shown that the strain background of an animal
   model can significantly modify the phenotypes of neurodegenerative
   diseases, such as lysosomal storage disorders and AD [[246]27,
   [247]81]. Investigating the phenotypes caused by PGRN deficiency in
   mice with different backgrounds is not only valuable for understanding
   the mechanisms underlying disease onset and progression but also
   important for developing proper disease mouse models. In this study, we
   found that mouse strain background modifies PGRN deficiency phenotypes.
   Specifically, PGRN deficiency in the FVB mouse strain leads to an
   earlier onset and more pronounced phenotypes, including gliosis,
   lysosomal defects, and TDP-43 pathology, as well as more significant
   alterations of gene and protein expression, compared to the B6 mouse
   strain (Fig. [248]1-Figs. 5 and [249]10, and Fig. [250]11). Our studies
   suggest that the Grn^−/− FVB mouse strain could be a valuable tool for
   studying the effects of PGRN loss, particularly for identifying early
   events involved in the development of PGRN deficiency phenotypes.

   RNA-seq and proteomics analysis revealed significant differences
   between the two mouse strains, with PGRN loss leading to distinct
   alterations in gene and protein expression as well as in various
   biological pathways (Figs. [251]10 and [252]11). These analyses also
   identified common proteins regulated by PGRN, such as SCRB2, PPT1,
   SMPD1, and ASAH1, which have been shown to be affected by PGRN
   deficiency in previous studies [[253]47, [254]82, [255]83]. Moreover,
   RNA-seq analysis suggested that lysosomal and immune pathways are
   upregulated by PGRN deficiency in both strains (Fig. [256]10h), while
   proteomics analysis suggests that inorganic transmembrane transport is
   upregulated in both strains (Fig. [257]11g), underscoring the role of
   PGRN in these pathways. Interestingly, both RNA-seq and proteomic
   analysis revealed the differential regulation of mitochondrial pathways
   by PGRN in the two strains (Figs. [258]10h and i and [259]11g and h).

Connection between mitochondrial dysfunction and PGRN deficiency phenotypes

   Mitochondrial dysfunction, including impaired ATP production, increased
   oxidative stress, and defective mitophagy, has been closely associated
   with neurodegenerative diseases, including LSDs and FTLD
   [[260]84–[261]87]. In our study, we observed that PGRN loss affects
   mitochondrial function differently in FVB and B6 strain backgrounds
   (Fig. [262]10h and i, and Fig. [263]11g and h). In response to PGRN
   loss, mitochondrial function is likely to be increased in the B6 mice,
   as evidenced by the upregulation of the oxidative phosphorylation/ATP
   metabolic pathway, which might exert a protective effect. In contrast,
   this compensatory response is absent in FVB mice, which might explain
   the more severe lysosomal and inflammatory phenotypes observed. These
   findings indicate that impaired mitochondrial function might enhance
   the PGRN deficiency phenotypes.

   A key pathological feature observed in PGRN-deficient mouse brains is
   the accumulation of lipofuscin in microglia and neurons (Fig.[264]S2).
   While the underlying mechanism behind this remains unclear, growing
   evidence points to a significant role of mitochondria in lipofuscin
   accumulation [[265]88–[266]92]. Intriguingly, several studies have
   demonstrated that PGRN regulates various aspects of mitochondrial
   functions. Disrupted mitochondrial integrity, impaired mitochondrial
   biogenesis, or defective mitophagy have been observed in conditions of
   PGRN deficiency [[267]93–[268]95]. Thus, it is plausible that
   mitochondrial dysfunction contributes to impaired lysosomal activity
   and lipofuscin accumulation in the PGRN-deficient mouse brain, which
   might lead to increased TDP-43 aggregation (Fig. [269]5) since the
   autophagy-lysosomal pathway is critical for TDP-43 turnover
   [[270]96–[271]98]. Moreover, mitochondrial defects may trigger
   inflammation [[272]99, [273]100]. Interestingly, a recent study has
   demonstrated that PGRN deficiency leads to mitochondrial hyperfusion
   and bioenergetic deficits in mice, which activate complement C3a-C3aR
   signaling, subsequently triggering microglia activation [[274]101].
   Therefore, mitochondrial dysfunction might contribute to both lysosomal
   defects and enhanced inflammation observed under PGRN deficiency
   conditions, and future studies are needed to dissect the detailed
   mechanisms.

Electronic supplementary material

   Below is the link to the electronic supplementary material.
   [275]Supplementary Material 1^ (2MB, pdf)

Acknowledgements