Complement receptor 1 is expressed on brain cells and in the human brain

Abstract Genome wide association studies (GWAS) have highlighted the importance of the complement cascade in pathogenesis of Alzheimer's disease (AD). Complement receptor 1 (CR1; CD35) is among the top GWAS hits. The long variant of CR1 is associated with increased risk for AD; however, roles of CR1 in brain health and disease are poorly understood. A critical confounder is that brain expression of CR1 is controversial; failure to demonstrate brain expression has provoked the suggestion that peripherally expressed CR1 influences AD risk. We took a multi‐pronged approach to establish whether CR1 is expressed in brain. Expression of CR1 at the protein and mRNA level was assessed in human microglial lines, induced pluripotent stem cell (iPSC)‐derived microglia from two sources and brain tissue from AD and control donors. CR1 protein was detected in microglial lines and iPSC‐derived microglia expressing different CR1 variants when immunostained with a validated panel of CR1‐specific antibodies; cell extracts were positive for CR1 protein and mRNA. CR1 protein was detected in control and AD brains, co‐localizing with astrocytes and microglia, and expression was significantly increased in AD compared to controls. CR1 mRNA expression was detected in all AD and control brain samples tested; expression was significantly increased in AD. The data unequivocally demonstrate that the CR1 transcript and protein are expressed in human microglia ex vivo and on microglia and astrocytes in situ in the human brain; the findings support the hypothesis that CR1 variants affect AD risk by directly impacting glial functions.

CR1 is a membrane receptor that controls complement activation by accelerating decay of the C3 convertase and acting as a co-factor for factor I (FI)-mediated conversion of C3b to iC3b, the ligand for the phagocytosis receptor complement receptor 3 (CR3; CD11b/CD18) (Jensen et al., 2021). Soluble CR1 (sCR1) is a disease biomarker in systemic lupus erythematosus (Arora et al., 2004;Hamer et al., 1998) and recombinant sCR1 was developed for therapy with some success in animal models (Carpanini et al., 2019;Morgan & Harris, 2015). The CR1 gene, located in the "regulators of complement activation" gene cluster on chromosome 1q32, encodes a type 1 transmembrane protein comprising a chain of 60 amino acid repeating units (short consensus repeats; SCR), grouped into sets of seven (long homologous repeats; LHR), each a functional unit with different complement binding properties (Figure 1). Four co-dominant CR1 alleles exist, differing in LHR number. The CR1*1 variant (allele frequency 0.87) comprises four LHRs, while the CR1*2 variant (allele frequency 0.11) contains an additional LHR inserted between LHRs A and B, providing an extra C3b/C4b binding site ( Figure 1) (Krych-Goldberg & Atkinson, 2001). CR1*2 is robustly associated with risk for late-onset AD, faster cognitive decline and greater neuropathological burden (Schmidt et al., 2014;Torvell et al., 2021). CR1*2 carriers have lower levels of brain Aβ, suggesting a role in clearance (Thambisetty et al., 2013), a paradoxical observation prompting the suggestion that Aβ is redistributed into a more neurotoxic form (Gandy et al., 2013).
The cellular and molecular roles of CR1 in the brain and how it influences AD risk are poorly understood. Indeed, brain expression of CR1 remains controversial. A study of brain CR1 expression using western blots, quantitative real-time PCR (qRT-PCR), and immunohistochemistry concluded that CR1 was not expressed (Johansson et al., 2018), while others found no CR1 immunoreactivity in normal brain tissue using available reagents (Singhrao et al., 1999). In contrast, CR1 expression on astrocytes was detected in AD and control brains (Fonseca et al., 2016), and CR1 protein and/or mRNA expression has been reported in astrocytes (Gasque et al., 1996) and neurons (Hazrati et al., 2012). A recent study showed CR1 expression in human primary fetal microglia and in F I G U R E 1 Representation of CR1 structure and ligand binding sites. From the amino terminus (NH 2 ) CR1*1 comprises four long homologous repeats (LHRs) each composed of seven short consensus repeats (SCRs) of 60-70 amino acids, two additional SCRs, a transmembrane segment (TM) and an intracytoplasmic carboxy-terminal domain (IC-COOH). Each circular block represents an SCR (numbered 1-30). There are three C4b binding sites (SCR 1-3, 8-10, and 15-17) and two C3b binding sites (SCR 8-10 and 15-17). SCRs 22-28 bind C1q, MBL, and ficolins. CR1*2 has an additional LHR domain (LHR-S) and consequently an extra C3b/C4b binding site. Adapted from Torvell et al. (2021). microglia differentiated from human induced pluripotent stem cells (iPSC) (Haenseler et al., 2017). Three independent studies reported CR1 mRNA expression in AD cortical homogenates (Allen et al., 2015;Holton et al., 2013;Karch et al., 2012). These conflicting studies prompted us to adopt a multi-pronged approach to test CR1 expression in human microglia in situ and in vitro. The clarification is critical because the presence of CR1 on brain immune cells would support a direct role in brain immune and inflammatory processes that might explain the impact of CR1 on AD pathology.

| Chemicals
All reagents and tissue culture plastics, except where otherwise stated, were from Fisher Scientific (Loughborough, United Kingdom) or Sigma Aldrich (Gillingham, United Kingdom) and of analytical grade.

| Human tissue
EDTA blood samples (5 mL) were collected from consented healthy donors and anonymized. Erythrocyte membranes were isolated by standard methods (Zelek et al., 2019). Frozen post-mortem brain tissue from five AD cases (Braak VI) and five age and sex-matched controls were used in this study (Table S1). The primary auditory cortex (BA41/42) was chosen as a well-described site of early AD-associated pathology with strong pathology burden at end stage disease (Jackson et al., 2019).

| CR1 junction fragment analysis
Genomic DNA was extracted from EDTA blood using the E.Z.N.A. ® Tissue DNA Kit (Omega Bio-tek); PCR products were amplified from 100 ng genomic DNA in a final volume of 25 μL, with 12.5 μL 2Â GoTaq Green Master Mix, and 1 μL of each PCR primer at 10 μM.
The PCR primers were 5 0 -AAT GTG TTT TGA TTT CCC AAG ATC AG-3 0 (forward) and 5 0 -CTC AAC CTC CCA AAG GTG CTA-3 0 (reverse). A touch-down PCR protocol was used (Kucukkilic et al., 2018), with an initial denaturation step of 95 C for 2 min, followed by 20 cycles of 95 C for 30 s, 70 C for 30 s decreasing by 0.5 C every cycle to 60 C, and 70 C for 30 s. These 20 cycles were then followed by 15 cycles of 95 C for 30 s, 60 C for 30 s and 70 C for 30 s, then a final extension step of 70 C for 5 min. PCR products were analyzed on a 1.5% agarose gel stained with SYBR Safe DNA gel stain and visualized under ultraviolet light in a Syngene Gbox.

| CR1 density polymorphism genotyping
The density polymorphism of CR1 dictating expression level on erythrocytes was determined by identifying the HindIII restriction fragment length polymorphism (RFLP) using PCR amplification followed by restriction enzyme digestion (Cornillet et al., 1991). PCR products were ampli- England Biolabs) for 1 h at 37 C followed by inactivation for 20 min at 80 C. The digested products were then analyzed on a 2% agarose gel along with control samples for each genotype. The gel was stained with SYBR Safe DNA gel stain and visualized under ultraviolet light in a Syngene Gbox. HindIII digestion did not alter the PCR product (1.8 kb) from individuals who were homozygous for the CR1 high-density allele (HH), whereas the PCR product was fully cleaved to 1.3 and 0.5 kb bands in samples from individuals homozygous for the CR1 low-density allele (LL).

| Apolipoprotein E genotyping
Apolipoprotein E (APOE) status was determined as previously described (Ingelsson et al., 2003). PCR products were amplified from
Selected guides had one potential off-target binding site within CR1L.
Cas9 enzyme and guides in a ribonucleoprotein complex were introduced into KOLF2 cells by electroporation/nucleofection as described (Bruntraeger et al., 2019). Cells were plated onto 50 ng/cm 2 vitronectin (ThermoFisher)-coated 10 cm dishes in mTesR medium; after 7 days, single colonies were picked and plated across duplicate 96-well plates, one for genomic DNA extraction using the E.Z.N.A. ® Tissue DNA Kit (Omega Bio-tek), the other for clonal expansion.
Clones were screened by PCR of genomic DNA with edit-flanking primers and separated on 2% agarose gels (unedited clones 733 bp PCR product, edited clones 588 bp product) and confirmed by Sanger sequencing (Eurofins Genomics). Edit-positive clones were re-plated onto vitronectin-coated culture dishes, sub-colonies picked and expanded.
2.9 | iPSC differentiation to microglia and culture

| ACM production and characterization
iPSC were differentiated to astrocytes in ADF medium supplemented with ciliary neurotrophic factor as previously described (Maguire et al., 2021). ACM was generated from fully differentiated iPSCastrocytes grown in Nunc T-500 triple layer flasks (Thermo Fisher); ADF medium was removed and replaced with ADF supplemented with 1% vitamin A neurobrew (Miltenyi Biotech). Medium was collected 48 h later, replaced with fresh medium, and stored at À80 C.
This cycle was repeated through 6-8 harvests. All harvests were combined to generate a single batch of ACM, which was used for all experiments. Chemokine (C-C motif) ligand 2 (CCL2; also known as monocyte chemoattractant protein 1) was measured in ACM by ELISA anti-CR1 SCR1-3 mouse mAb 3E10 generated by immunization with recombinant CR1 SCRs 1-3 (Banz et al., 2007); two novel anti-CR1 mouse mAb, MBI35 and MBI38, generated in house using full-length sCR1 as immunogen and selected for high affinity binding of CR1. All detected CR1 in immunostaining and western blotting of erythrocytes and for all, specificity was confirmed by demonstrating that preincubation with excess full-length sCR1 ablated staining (Piddlesden et al., 1994).
Coverslips were washed, incubated with Alexa Fluor conjugated goat anti-rabbit or anti-mouse IgG (ThermoFisher) and Hoechst dye,

| Statistics
All values were expressed as mean ± SEM. Data were plotted using

| CR1 is expressed in KOLF2 iPSC-derived microglia and is not detected in CR1 knockout cells
CR1 expression was assessed in MPC and microglia derived from the human iPSC line, KOLF2, and from KOLF2 in which the CR1 gene had been knocked out (KO) by CRISPR gene editing (CRISPR KO strategy summarized in Figure S1). Three clones for KOLF2 WT and KO lines were differentiated into microglia via EB formation and characterized at the MPC and iPSC-microglia stages of differentiation for cell-specific protein markers expression and ability to phagocytose bioparticles ( Figure S2). Abundant surface expression of CR1 protein, co-staining with CD11b, was seen on WT iPSC-microglia stained with anti-CR1 while KOLF2 CR1 KO derived microglia, stained with CD11b, were negative, confirming that CR1 was deleted and further demonstrating specificity of the selected antibodies for the CR1 protein ( Figure 3a).

Expression of CR1 on KOLF2-derived microglia and absence in KO lines
was also demonstrated by western blotting and qRT-PCR (Figure 3b,c).
Western blot revealed that KOLF2 cells were CR1*1 homozygous. SDS-PAGE of lysates run in parallel and stained with Coomassie confirmed consistent protein loading. Erythrocyte CR1 density polymorphism was determined by RFLP analysis in 15 donors selected for further study; 10 donors were highdensity allele homozygous (HH) and 5 were heterozygous (HL) ( Figure S3c). Due to its status as a strong genetic risk factor for AD (Saunders et al., 1993), donor APOE variant status was tested by RFLP analysis in 27 donors; 22 were ε3/ε3 ( Figure S3d). Screening data are summarized in Table S3.
Three clones for each homozygote and heterozygote line were selected for differentiation into microglia via EB formation ( Figure S5a). MPC were large, round cells with filopodia, expressing CD14, CD45, and CD11b ( Figure S5b). MPC were further differentiated to iPSC-derived microglia, confirmed by demonstrating expression of microglia-specific markers CX3CR1, TMEM119, IBA1, CD45, CD68, and CD11b ( Figure S5c). Acquisition of phagocytic capacity was confirmed by demonstrating phagocytosis of pHrodo™ E. Coli BioParticles ( Figure S5d). F I G U R E 2 CR1 expression in HMC3 and IMhu microglial cell lines. (a) Dual immunofluorescence staining of HMC3 (i-iii) and IMhu (iv-vi) cells with anti-CR1 antibodies: affinity-purified polyclonal anti-CR1 (i; iv), mAb MBI38 (ii) and mAb MBI35 (v) with microglial markers IBA1 (ii, iii, v, vi) or CD45 (i, iv). Impact of pre-adsorption with excess sCR1 on staining with mAb MBI38 (iii) and mAb MBI35 (vi) are shown. Nuclei are stained with Hoechst 33342 in the merged images for each set of plates. All images are captured at 40Â except set (iii) which are at 80Â; scale bar shown in right image for each set is 50 μm in all. (b) CR1 protein was detected in HMC3 and IMhu cells by Western blotting using all anti-CR1 antibodies (polyclonal shown as example). Protein lysate from red blood cells (RBC) from a CR1*1/CR1*2 donor was used as a positive control. (c) CR1 mRNA was detected by qRT-PCR using intron-spanning primers (Table S2) in HMC3, IMhu, THP-1 (human monocytic cell line) but not SH-SY5Y (neuronal) cells. The reaction without reverse transcriptase (no RT) was used as a negative control.

| CR1 is expressed on microglia and astrocytes in human brain tissue
Frozen post-mortem brain samples (region BA41/42) obtained from five AD cases (Braak VI) and five age and sex-matched controls (Table S1) (Table S1, Figure 5e,f). qRT-PCR analysis was performed on tissue-extracted RNA for assessing expression of CR1 and tissue/cell-specific markers. CR1 transcripts were detected in all control and AD whole brain extracts tested (Figure 6a).
Transcripts for microglial (IBA1), astrocyte (GFAP) and neuronal (RBFOX3) and an endogenously expressed control (SDHA) were also detected ( Figure 6a). CR1 mRNA expression was significantly increased ($5-fold) in the AD samples compared to controls ( Figure 6b). Expression of mRNA for the microglial marker IBA1 and the astrocytic marker GFAP were also significantly increased in AD compared to control samples, indicative of pathology-associated microgliosis and astrogliosis; in contrast, expression of mRNA for the neuronal marker, RBFOX3, was similar in control and AD samples ( Figure 6c).
F I G U R E 3 CR1 expression in KOLF2-derived microglia. (a) CR1 expression in iPSC-microglia in WT KOLF2-derived microglia shown by immunofluorescence. CR1 was expressed in a granular, membrane-associated pattern in KOLF2-derived microglia but was not detected in CR1 KO KOLF2. Cells are co-stained with microglial marker CD11b. Maximum projections of Z-stacks are presented. Scale bar: 30 μm.
(b) CR1 protein was detected in WT iPSCmicroglia by western blotting with the different anti-CR1 antibodies (polyclonal shown here). Protein lysate from RBC from a CR1*1/CR1*2 donor was used as positive control. CR1 KO cells showed trace staining. (c) CR1 transcript was quantified by qRT-PCR using intron-spanning primers (Table S2) in WT and CR1 KO KOLF2 iPSC and iPSC-microglia. Histogram shows the relative mRNA expression normalized to nondifferentiated cells (iPSC); **p < .01.

| DISCUSSION
Over the last 20 years, an abundance of evidence has implicated complement in AD pathogenesis (reviewed in Morgan, 2018). GWAS have identified AD risk single nucleotide polymorphisms in complement genes (reviewed in Torvell et al., 2021), including CR1, CLU, and more recently a suggestive association in C1S (Bellenguez et al., 2022). Biomarker studies have also identified alterations in complement proteins and activation products in blood and/or cerebrospinal fluid that distinguish controls from mild cognitive impairment and are predictive of progression to AD (Hakobyan et al., 2016;Morgan et al., 2019). In post-mortem immunohistochemical studies of AD brain, C1q, C4b, C3b/iC3b, and membrane attack complex (MAC) have all been shown to co-localize with plaques and tangles (Ishii & Haga, 1984;Rogers et al., 1992;Veerhuis et al., 2003). Data from animal models have implicated complement in amyloid clearance and in synapse loss (Maier et al., 2008;Shi et al., 2017;Wyss-Coray et al., 2002); notably, our recent study demonstrated that MAC formation is an important driver of synapse loss in AD models (Carpanini et al., 2022).
Although long established among the top AD-associated GWAS hits, the specific cellular and molecular roles of CR1 in brain health and disease are poorly understood. Improved understanding of the expression and roles of CR1 in the brain in health and disease is needed to explain its association with AD and to facilitate rational design of diagnostic or therapeutic tools. Complement inhibition is a proven treatment in numerous diseases, including the neurological diseases, neuromyelitis optica and myasthenia gravis (Huda et al., 2014;Pittock et al., 2019). In neurodegeneration, anticomplement drugs are in clinical trials for treatment of Huntington's disease (NCT04514367) and amyotrophic lateral sclerosis F I G U R E 4 CR1 expression in donor-derived iPSC-microglia. (a) CR1 expression in MPC and iPSC-microglia is shown by immunofluorescence. Cells showed membrane-associated staining with the pAb against CR1 and the anti-CR1 mAbs, MBI38, and 3E10; sCR1 pre-adsorption ablated staining for each antibody (MBI38 shown as example). Maximum projections of Z-stacks are presented. Representative examples of lines homozygous for CR1*1 and CR1*2 variants are shown. Scale bar: 50 μm. (b) CR1 protein was detected in MPC and iPSC-microglia expressing the CR1*1 and CR1*2 variants by western blotting with the different anti-CR1 antibodies (polyclonal anti-CR1 shown as example). Protein lysate from RBC from a CR1*1/CR1*2 donor was used as a positive control and anti-tubulin as the loading control. (c) CR1 transcripts were detected by qRT-PCR using intron-crossing primers introns (Table S2) in iPSC (1), MPC (2), and iPSC-microglia (3) expressing the CR1*1 and CR1*2 variants. Histograms show the relative mRNA expression normalized to non-differentiated cells (iPSC). *p < .05; **p < .01; ***p < .001.
F I G U R E 5 CR1 protein expression in the human brain. Immunofluorescent staining of frozen sections of control (a) and AD (b) human brain tissue ( Expression of CR1 in the brain has been the subject of contrasting data and debate and published reports of CR1 protein and message in the brain and on brain cells are confusing and contradictory. In early studies, protein expression was variously reported as only in astrocytes (Gasque et al., 1996), restricted to phagocytic Kolmer cells of the choroid plexus and ependymal cells (Canova et al., 2006), no expression at all (Singhrao et al., 1999), or on choroid plexus, microglia and neurons in AD and control brain tissue (Hazrati et al., 2012). A comprehensive analysis of CR1 staining tested a panel of seven mAbs and two antisera from various sources in formalin-fixed AD and control brain (Fonseca et al., 2016); while most of these reagents did not stain the tissue, two mAbs (8C9.1 and J3B11) from the panel specifically stained astrocytes and specificity was confirmed by preadsorption with sCR1 and staining of isolated astrocytes. In contrast, a subsequent report from the same investigators tested four mAbs against CR1, not including the two shown in their earlier publication, and found no specific staining in brain parenchyma although vasculature was stained (Johansson et al., 2018). CR1 mRNA expression in AD and control brain has been widely reported, including the demonstration that CR1 expression levels were increased in AD brain and associated with cognitive score (Karch et al., 2012) and that CR1 gene expression levels strongly associated with AD risk (Allen et al., 2015).
Microarray data from the Allen Human Brain Atlas also demonstrates significant CR1 expression in brain (Hawrylycz et al., 2012). A recent report described CR1 mRNA expression in fetal and iPSC-derived microglia (Haenseler et al., 2017). anti-CR1 antibodies and controls to demonstrate expression of CR1 in situ in the human brain. Specific staining was seen in control and AD brain sections; numbers of CR1-positive cells were $5-fold higher in AD GM regions compared to non-AD controls. Double-staining with cell-type specific markers confirmed that not only microglia but also astrocytes were CR1-positive, the latter particularly in the AD brain.
Of note, neurons (stained with HuC/D) were negative in all samples.
CR1 transcript was also detected in both control and AD human brain tissue extracts and expression was $6-fold higher in AD tissue compared to control.
We speculate that failure to detect CR1 protein expression in some studies may be the result of tissue selection and processing (we used frozen tissue and optimized post-fixation conditions), choice of antibodies (we used well-characterized mAbs raised against fulllength sCR1 pre-selected for high binding affinity and an affinitypurified polyclonal) and lack of relevant controls. The same factors are F I G U R E 6 CR1 mRNA expression in the human brain. (a) Agarose gel demonstrating detection of CR1 transcripts in control and AD human brain tissue using specific primers. Transcripts for the cell type-specific markers IBA1, GFAP, and RBFOX3 and the housekeeping gene SDHA were also present in the brain extracts. (b) Relative CR1, IBA1, GFAP and RBFOX3 mRNA expression (-fold) in control and AD human brain tissue; CR1 mRNA showing $5-fold greater expression in AD brain. **p < .01; ***p < .001, data from five AD and five control brains.
likely to be responsible for the different cell specificities in those studies that do report CR1 protein expression in brain.
Taken together, our multisource data incontrovertibly demonstrate expression of CR1 message and protein in glial cells and in the brain. Our initial focus was on microglia, but the demonstration that astrocytes also express CR1 opens new avenues of work, particularly relevant given the critical roles of astrocytes in synaptic elimination (Lee et al., 2021). Our ambition is to explain how the CR1*2 variant confers AD risk, and to this end, we have generated CR1*1/CR1*1, CR1*1/CR1*2, and CR1*2/CR1*2 iPSC lines. These iPSC lines provide us with essential tools to study the functional differences between risk and non-risk variants that cause the association with AD risk. Here we show that all variant combinations were expressed and membrane-localized on iPSC microglia. Our initial characterization suggests reduced CR1 protein expression in CR1*2/CR1*2 compared to CR1*1/CR1*1 iPSC microglial lines despite increased CR1 mRNA in the former. This interesting disconnect is the subject of ongoing work that will extend the expression analysis to define precisely how the variants affect CR1 levels and distribution on iPSCderived microglia and astrocytes and explore the impact of the CR1 variants on C3 fragment processing and phagocytic capacity for relevant targets. Detailed understanding of the impact of CR1 variants on interactions with its partners in the complement system and elsewhere will inform disease mechanism and signpost routes to novel disease modifying therapies. All authors read and approved the final manuscript.