The relation between BTK expression and iron accumulation of myeloid cells in multiple sclerosis

Abstract

Activation of Bruton's tyrosine kinase (BTK) has been shown to play a crucial role in the proinflammatory response of B cells and myeloid cells upon engagement with B cell, Fc, Toll‐like receptor, and distinct chemokine receptors. Previous reports suggest BTK actively contributes to the pathogenesis of multiple sclerosis (MS). The BTK inhibitor Evobrutinib has been shown to reduce the numbers of gadolinium‐enhancing lesions and relapses in relapsing–remitting MS patients. In vitro, BTK inhibition resulted in reduced phagocytic activity and modulated BTK‐dependent inflammatory signaling of microglia and macrophages. Here, we investigated the protein expression of BTK and CD68 as well as iron accumulation in postmortem control (n = 10) and MS (n = 23) brain tissue, focusing on microglia and macrophages. MS cases encompassed active, chronic active, and inactive lesions. BTK⁺ and iron⁺ cells positively correlated across all regions of interests and, along with CD68, revealed highest numbers in the center of active and at the rim of chronic active lesions. We then studied the effect of BTK inhibition in the human immortalized microglia‐like HMC3 cell line in vitro. In particular, we loaded HMC3 cells with iron‐dextran and subsequently administered the BTK inhibitor Evobrutinib. Iron treatment alone induced a proinflammatory phenotype and increased the expression of iron importers as well as the intracellular iron storage protein ferritin light chain (FTL). BTK inhibition of iron‐laden cells dampened the expression of microglia‐related inflammatory genes as well as iron‐importers, whereas the iron‐exporter ferroportin was upregulated. Our data suggest that BTK inhibition not only dampens the proinflammatory response but also reduces iron import and storage in activated microglia and macrophages with possible implications on microglial iron accumulation in chronic active lesions in MS.

This study investigates the impact of Bruton's tyrosine kinase (BTK) activation in multiple sclerosis (MS), finding a positive correlation between iron accumulation and BTK expression in myeloid cells in post‐mortem MS brain tissue. We further demonstrate that the BTK inhibitor Evobrutinib modulates iron levels together with the inflammatory response in human microglia‐like cell clone 3 (HMC3) cells. Overall, the findings suggest that BTK inhibition not only mitigates pro‐inflammatory activation but also influences iron import and storage of microglia, potentially impacting myeloid cell iron accumulation in chronic active MS lesions.

1. INTRODUCTION

Multiple sclerosis (MS) is an inflammatory disease of the central nervous system, leading to the formation of focal demyelinating lesions and axonal as well as neuronal degeneration [1, 2]. In early MS, demyelinating activity is mainly found in active lesions, which show considerable but transient axonal injury across the lesion areas [3]. In chronic active lesions, sustained inflammation and axonal injury are concentrated at the edges of the lesions [4]. The inflammation in these chronic lesion rims is mainly composed of HLA‐DR⁺ microglia and macrophages, few parenchymal T cells, and even fewer B cells, with myeloid cell infiltration and axonal transections being spatially related [1, 3]. Rim‐related myeloid cells were shown to accumulate iron [5, 6] forming iron rims, which are visible by iron‐sensitive magnetic resonance imaging as paramagnetic rim lesions (PRLs) [7]. These PRLs indicate myeloid cell‐dominated rim inflammation and chronic lesion activity in vivo [8], which allowed to clarify the dynamics of these chronic active lesions over time [9] and their relation to a more severe MS disease course [10, 11]. Although paramagnetic rims (PRs) typically remain stable for several years, the rim signal is attenuated over time and eventually lost after a time period of about 7 years or longer [9, 12]. It seems reasonable that the disappearance of the PR indicates suspended chronic lesion activity, since inactive or remyelinated lesions do not show profound rim‐related iron accumulation [8]. Thus, PRs are already used to monitor chronic activity of lesions and have the potential to reflect progression independent of relapse activity in MS patients. PRs have been shown to be attenuated upon treatment with dimethyl fumarate when compared with glatiramer acetate [13]. Inhibition of Bruton's tyrosine kinase (BTK) is currently intensely studied as a therapeutic option in MS due to its combined effects on B and myeloid cells as well as the good blood–brain‐barrier penetration of BTK inhibitors [14, 15, 16]. BTK is a receptor‐free tyrosine kinase crucial for B cell functioning [17]. In addition, it is involved in innate immune processes due to its expression by, for example, neutrophilic granulocytes [18] and myeloid cells [19]. In the latter, it signals downstream of Toll‐like receptors (TLRs), leading to the activation of the NLR family pyrin domain containing 3 (NLRP3) inflammasome and subsequent cytokine release [20]. In this respect, a study reported a link between TLR7 activation and myeloid cell heme iron uptake via Low‐density lipoprotein receptor‐related protein 1 (LRP1) induction by BTK signaling [21]. Another study reported BTK‐dependent induction of heme oxygenase 1 (HO1) upon TLR4 activation in primary mouse alveolar macrophages [19]. Thus, previous studies have related BTK activation to heme iron uptake of myeloid cells and subsequent BTK‐mediated upregulation of HO1. Heme contains ferrous iron, which is potentially harmful due to its high pro‐oxidative capacity, particularly upon extracellular liberation [22]. Recently, we have shown that hemoglobin and its component heme are potentially important iron sources for iron‐laden myeloid cells at the rims of chronic active lesions in MS [23]. This iron was presumably liberated intracellularly from heme by HO1 and then stored within ferritin in the myeloid cells. Moreover, we found hepcidin upregulation in these iron‐laden myeloid cells. Hepcidin acts as a functional antagonist of the cellular iron exporter ferroportin (FPN), thus mediating myeloid cell iron retention, particularly under proinflammatory conditions [24]. Several previous studies have shown the proinflammatory modulation of myeloid cells upon their iron loading in various pathological and experimental circumstances [25, 26, 27, 28], suggesting that iron accumulation might augment the proinflammatory myeloid cell phenotype observed at chronic active lesion rims. Here, we aimed at elucidating the expression of BTK in histopathological sections of MS and the effect of BTK inhibition on myeloid cell iron uptake and expression of several related iron‐handling and inflammation‐related proteins in vitro, including iron importers, ferritin, and NLRP3.

2. MATERIALS AND METHODS

2.1. Patient samples

Postmortem brain tissue from MS patients was collected from the archives of the Division of Neuropathology and Neurochemistry, Department of Neurology, Medical University of Vienna, Austria, and the Department of Neuroimmunology at the Center for Brain Research, Medical University of Vienna, Austria. The study cohort comprised autopsy brain tissue blocks from 10 controls with neither a history of neurological disease nor evidence of brain lesions, and 23 MS cases. No statistical difference was observed regarding the average ages of the MS and control cohorts (p = 0.1558). Basic epidemiological and clinical information on the cases can be found in Table 1. The neuropathological part of the study was approved by the ethics committee of the Medical University of Vienna (535/2004/2016).

TABLE 1.

Clinical data of study cohort.

Case	Age, years, median	Sex, F:M ratio	Disease duration, months	Lesion types
Controls 1–10	63.50	6:4	‐
MS 1–23	54.00	13:10
MS 1	76	f	na	1× ia
MS 2	56	f	na	1× ia
MS 3	65	m	300	1× ia
MS 4	73	m	na	1× ia
MS 5	78	f	312	1× ia
MS 6	44	f	na	1× ia, 1× ca, 1× a
MS 7	62	f	144	1× ia
MS 8	61	f	na	1× ia, 1× ca
MS 9	52	m	288	1× ia, 1× ca
MS 10	46	f	444	1× ia
MS 11	50	f	>120	1× ca
MS 12	46	f	na	1× ca
MS 13	67	m	87	1× ca
MS 14	53	m	168	1× ca
MS 15	36	m	60	1× ca
MS 16	53	f	240	1× ca
MS 17	34	m	>120	1× ca
MS 18	37	f	7	1× a
MS 19	69	f	2	1× a
MS 20	28	m	6	1× a
MS 21	64	m	6	1× a
MS 22	59	f	444	1× a
MS 23	34	m	24	1× a

Abbreviations: a, active; ca, chronic active; f, female; ia, inactive; m, male; MS, multiple sclerosis; na, not available.

2.2. Neuropathology

Basic tissue characterization of controls and MS cases was performed using sections routinely stained for hematoxylin & eosin. MS lesion staging was accomplished on Klüver‐Barrera‐PAS stainings and immunohistochemistry (IHC) for the phagocytosis marker CD68 for microglia and macrophages. Active MS lesions (n = 7) were characterized by a hypercellular lesion area densely populated by macrophages, many of them containing Luxol Fast Blue‐positive myelin degradation products. Chronic active lesions (n = 10) displayed a hypocellular, completely demyelinated, inactive lesion core and a rim of microglia and macrophages. Inactive lesions (n = 10) showed an inactive lesion area without noticeable inflammation at their edges.

2.3. Immunohistochemistry

IHC was performed on formalin‐fixed, paraffin‐embedded human brain tissue. Tissue was cut into 3‐μm thick sections and mounted on glass slides. Sections were deparaffinized in descending concentrations of ethanol until distilled water. Endogenous peroxidase was blocked with 3% H₂O₂ in methanol for 10 min, and slides were rinsed with distilled water. Antigen retrieval was performed in a household food steamer (Braun) in Citrate buffer, pH 6. Slides were cooled to room temperature and rinsed three times with Tris‐buffered saline (TBS). Unspecific binding was blocked with 10% fetal calf serum (FCS)/DAKO diluent for 10 min. Primary antibodies were diluted in DAKO diluent and samples were incubated overnight at 4°C. The next day, slides were rinsed three times with TBS. For antibody detection, the secondary system Dako REAL EnVision Detection System (Agilent, K5007) was applied for 25 min. Stainings were developed with 3,3′‐diaminobenzidin (DAB) for 10 min. For counterstaining, sections were incubated in Mayer's hemalaun for 30 s, rinsed with tap water, and differentiated in 0.45% hydrochloric acid diluted in 70% ethanol. Sections were blued in warm tap water for 3 min and dehydrated in increasing concentrations of ethanol. Samples were mounted with Epredia™ Shandon™ Consul‐Mount™ (Fisher Scientific, 12590017).

The following primary antibodies were used: CD68 (Dako, GA60961‐2; 1:5000), BTK (abcam, ab208937; 1:200), Tim‐1 (Invitrogen, PA5‐98302; 1:2000), Lrp‐1 (abcam, ab92544; 1:500), and NLRP3 (Dako, EPR23094‐1; 1:200). For the primary antibody against BTK, Tyramide Signal Amplification system was used including a biotinylated secondary antibody (Dako, K0675), streptavidin‐conjugated horseradish peroxidase (Dako, K0675) and catalyzed signal amplification (CSA) [29] produced by the routine staff of the Division of Neuropathology and Neurochemistry, Medical University of Vienna.

2.3.1. Turnbull blue staining

Deparaffinization was performed as previously described. Sections were then incubated in 10% ammonium sulfide solution for 1.5 h and subsequently rinsed three times with distilled water. Then, slides were incubated in a 1:1 mixture of 20% potassium hexacyanoferrate (III) and 1% hydrochloric acid for 15 min, and sections were again washed three times with distilled water. Next, sections were incubated in a mixture of 0.01 M sodium azide (65 g/mol) in methanol with 30% H₂O₂ for 1 h and, subsequently, washed with 0.1 M Sorensen's phosphate buffer three times. For development, sections were incubated in a 1:50 mixture of DAB in 0.1 M Sorensen's phosphate buffer for 20 min. The reaction was stopped with tap water. Counterstaining and mounting were performed as described above.

2.3.2. Fluorescence triple staining

Deparaffinization, antigen retrieval, blocking of endogenous peroxidase, and blocking with 10% FCS were performed as previously described. The primary antibody targeting BTK (abcam, ab208937; 1:50) was applied overnight at 4°C. On day 2, the slide was washed with TBS and next, tyramide signal amplification system was used as described above. Upon that, the slide was again steamed for 30 min (pH 6) and washed with TBS. The primary antibodies against CD68 (Dako, GA60961‐2; 1:500) and ferritin light chain (FTL) (ProteinTech, 10727‐1‐AP; 1:1000) were then incubated overnight at 4°C. On day 3, the slide was washed with TBS and secondary systems were applied for 1 h at room temperature (RT): 488‐donkey‐anti‐mouse (Jackson Immunoresearch, AB_2338852; 1:800), Streptavidin‐Cy3 (Jackson Immunoresearch, AB_2337244; 1:1000) and Cy5‐donkey anti‐rabbit (Jackson Immunoresearch, AB_2338013; 1:1000). The slide was then washed with TBS and next, 4′,6‐diamidino‐2‐phenylindole (DAPI) staining was performed. Subsequently, slides were mounted with Prolong™ Gold Antifade Mountant (Invitrogen™, P36934) and dried for 24 h at 4°C in the fridge.

2.4. Quantification of cells

For the quantification of myeloid cell numbers, we evaluated different lesion types of MS brain tissues. Positive cells were counted manually within multiple regions of interest using the software NDP.view2 (Hamamatsu). Areas of interest comprised a size range from 0.05 to 0.9 mm². Counted cells were normalized to cells per mm² and used for statistical analysis.

2.5. Cell culture and treatments

Human microglia‐like clone 3 (HMC3) cells were obtained from the Institute of Science and Technology, Austria. Cells were cultured in Eagle's Minimum Essential Media (EMEM) (ATCC, 30‐2003) supplemented with 10% FCS (Sigma‐Aldrich, F9665) and 1% penicillin–streptomycin solution (Gibco, 15140122) in T‐75 flasks (Thermo Scientific™, 156499). Cells were maintained in an incubator at 37°C with 5% CO₂. For the experiments, cells from passages 5–10 were used. When cultures reached a confluency of 90%, cells were trypsinized, counted, and seeded into 12‐well plates (Corning, 3513) at a density of 0.1 × 10⁶ (2.8 × 10⁴ cells/mm²) and 60 mm dishes (TPP, 93060) at a density of 0.8 × 10⁶ (3.7 × 10⁴ cells/mm²).

Cells were washed with 1× Dulbecco's phosphate‐buffered saline (DPBS) (Thermo Scientific™, 14190094) and starved with Roswell Park Memorial Institute (RPMI) medium (Thermo Scientific™, 21875034) for 2 h to get rid of intracellular iron sources. Group 1 was cultured under homeostatic conditions in EMEM medium and used as controls. Group 2 was treated with 1 μg/mL iron‐dextran solution (Merck, D8517) for 20 h for iron loading followed by 12 h of incubation with RPMI only. Group 3 received iron‐dextran followed by treatment with 10 μM Evobrutinib (Selleckchem, S8777) in RPMI medium for 12 h to inhibit BTK in iron‐laden cells.

2.6. RNA isolation, reverse transcription, and quantitative polymerase chain reaction

Cells of 12‐well plates were washed with prewarmed 1× DPBS and lysed with TRK lysis buffer (VWR Life Science, VWRC13‐TRK‐02_P) complemented with 2% mercaptoethanol. Total RNA isolation and purification were performed with MicroSpin Total RNA Kit (VWR Life Science, 13‐6831‐02) according to the manufacturer's instructions. RNA concentration was measured with NanoDrop™ Lite spectrophotometer (Thermo Scientific™, ND‐LITE‐PR) and stored at −80°C until further processing. For cDNA generation, 600 ng RNA in a 20 μL reaction volume was reversely transcribed with the iScript™ cDNA Synthesis Kit (Bio‐Rad, 1708891) according to the manufacturer's instructions. For one quantitative polymerase chain reaction (qPCR) reaction a total volume of 10 μL consisting of 5 μL SsoAdvanced™ Universal SYBR® Green Supermix, 0.2 μL of forward and reverse primer, 3.6 μL water, and 1 μL of cDNA. For detection of BTK, the PrimePCR™ SYBR® Green Assay (Biorad, 10025636) was used.

All other primers were purchased from Sigma‐Aldrich (Table 2). The qPCR reaction was performed in the Applied Biosystems® 7900HT Fast Real‐Time PCR System. Amplification was performed in 40 cycles starting with initial denaturation step at 95°C for 45 s, followed by primer annealing and elongation at 60°C for 30 s. Amplicon specificity was determined with melt curve analysis. All reactions were performed in duplicates from at least three independent experiments. CT values were normalized to succinate dehydrogenase complex flavoprotein subunit A and glycerinaldehyde‐3‐phosphate‐dehydrogenase. Analysis of relative gene expression changes was calculated with the 2^−ΔΔCT method.

TABLE 2.

Primer sequences used for qPCR.

Gene symbol	Gene name	Primer sequence forward and reverse
SDHA	Succinate dehydrogenase complex subunit A	AAGGTGCGGATTGATGAGTA
		TTTGTCGATCACGGGTCTAT
GAPDH	Glycerinaldehyde‐3‐phosphate dehydrogenase	ATATTGTTGCCATCAATGACCC
		ATGACAAGCTTCCCGTTCTC
IBA1	Ionized calcium‐binding adapter molecule 1	TGAGCCAAACCAGGGATTTA
		CGTCTAGGAATTGCTTGTTGA
P2RY12	Purinergic receptor P2Y12	AGTCCCCAGGAAAAAGGTG
		GTTTGGCTCAGGGTGTAAGG
TLR4	Toll‐like receptor 4	ACTTGGACCTTTCCAGCAAC
		TTTAAATGCACCTGGTTGGA
NFKB1	Nuclear factor kappa B p50 subunit	TGAGTCCTGCTCCTTCCAA
		CTTCGGTGTAGCCCATTTGT
RELA	Nuclear factor kappa B p65 subunit	TAGGAAAGGACTGCCGGGAT
		CCGCTTCTTCACACACTGGA
NLRP3	NLR family pyrin domain containing 3	ATGAGTGCTGCTTCGACATC
		TTGTCACTCAGGTCCAGCTC
IL6	Interleukin 6	GGCACTGGCAGAAAACAACC
		GCAAGTCTCCTCATTGAATCC
IL1B	Interleukin1 beta	TACAGCTGGAGAGTGTAGATC
		CAAATTCCAGCTTGTTATTG
FTL	Ferritin light chain	CAGCCTGGTCAATTTGTACCT
		CGGTCGAAATAGAAGCCCAGAG
HAMP	Hepcidin	CTGACCAGTGGCTCTGTTTTCC
		AAGTGGGTGTCTCGCCTCCTTC
DMT1	Divalent metal transporter 1	TCTGGAGATCATGGGGAGTC
		TCCTCCTCAGGAATGGAGAT
TIM1	T‐cell immunoglobulin and mucin domain 1	AACTGTCTCTACCTTTGTTCCTCC
		GTTCTCTCCTTATTGCTCCCTG
FPN	Ferroportin	TCTTTGCTTGCGGTCCTGAT
		GAGCAAAACACCCAGCCATT
LRP1	Low density lipoprotein receptor‐related protein 1	CAACGGCATCTCAGTGGACTAC
		TGTTGCTGGACAGAACCACCTC

2.7. Protein isolation and Western blot

Cells in 60‐mm dishes were washed twice with prewarmed PBS and incubated with ice‐cold PBS for 5 min. For the following steps, dishes were kept on ice. Next, cells were incubated with RIPA buffer (Thermo Scientific™, 89901) mixed with 0.1% Halt™ Protease and Phosphatase Inhibitor Cocktail (Thermo Scientific™, 78440) at 4°C on a rocking shaker for 20 min. Subsequently, cells were scraped with a cell scraper, and the lysate was collected in a centrifuge tube. The lysate was then vortexed for 10 s and incubated on ice for further 20 min. The lysate was then centrifuged at 16,000× g for 10 min at 4°C. The supernatant was then transferred to a new centrifuge tube and stored at −20°C. Protein concentration was determined with Pierce™ BCA Protein Assay Kit (Thermo Scientific™, 23225) according to the manufacturer's instructions. Protein separation was performed with SDS‐PAGE in a 12% Tris‐Glycine buffer system. Proteins were transferred in a semi‐dry blot to a polyvinylidene difluoride membrane. Membrane was blocked in EveryBlot Blocking Buffer (Bio‐Rad, 12010020) for 5 min. The following antibodies were used: ferritin (Sigma‐Aldrich, F5012; 1:2000), Lrp‐1 (abcam, ab92544; 1:500), Dmt‐1 (Bioss, bs‐3577R; 1:500), hepcidin (Bioss, bs‐10348R; 1:1000), Tim‐1 (Invitrogen, PA5‐98302; 1:500), RelA (NFkB p65) (Cell Signaling Technology, 8242S; 1:500), IL‐1b (Invitrogen, P420B; 1:500), and TLR4 (Novus Biological, 76B357.1; 1:500). Beta‐actin (abcam, ab6276; 1:40,000) was used as a loading control. All primary antibodies were diluted in blocking buffer and incubated at 4°C on a rocking shaker overnight. The next day, the membrane was washed three times with 1× DPBS and incubated with anti‐mouse (Biorad, 1706516) or anti‐rabbit secondary antibody (Biorad, 1706515), coupled to horseradish peroxidase. Band visualization was performed with Clarity™ Western ECL Substrate (Bio‐Rad, 1705060). Band intensities were detected with the imaging system Fusion FX6 (Vilber).

2.8. Immunocytochemistry

For single fluorescent cell stainings, cells were seeded into an eight‐well Nunc™ Lab‐Tek™ Chamber Slide (Thermo Scientific™, 177445). When cells reached a confluency of 90%, cells were fixed with 4% PFA in 1× DPBS at room temperature for 15 min. Cells were then washed three times with 1× DPBS and permeabilized with 0.1% Triton X‐100 for 5 min at room temperature. Cells were blocked with Dako REAL antibody diluent (Dako, S2022‐30‐2) for 30 min at room temperature. The following primary antibodies were diluted in 10% Dako REAL antibody diluent/1× DPBS and incubated with cells overnight at 4°C: Iba‐1 (FUJIFILM Wako Pure Chemical Corporation, 019‐19741; 1:200), ferritin (Sigma‐Aldrich, F5012; 1:300), and FPN (NovusBio, NBP2‐75923; 1:300). The next day, primary antibodies were aspirated and cells were washed three times with 1× DPBS to remove unbound antibody. As secondary antibodies, we used the Alexa Fluor 488 suitable for the species mouse and rabbit (Invitrogen, A‐21202; A‐21206) in a concentration of 1:500 in 1× DPBS/10% FBS and incubated cells for 1 h at room temperature. Nuclear staining was performed with DAPI (ThermoScientific™, 62248; 1:1000), and F‐actin was stained with Alexa Fluor™ Plus phalloidin conjugates (Invitrogen™, R415; 1:1000) for 20 min at room temperature. Next, chambers were removed and slides were mounted with ProLong™ Gold Antifade Mountant (Invitrogen™, P36934). Stainings were imaged with Olympus BX63 fluorescence microscope.

2.9. Statistics

2.9.1. Immunohistochemistry

All samples were tested for normality. Unpaired t‐test was used to determine the presence of a significant age difference between MS patient samples and control samples. Statistical analysis of cell countings was performed with GraphPad Prism 8. Correlations between two metric variables were computed using regular linear regression. Comparisons of cell quantifications of different regions of interests (ROIs) were performed with multiple t‐tests with Bonferroni Dunn correction. Statistical p values < 0.05 were regarded as significant.

2.9.2. Cell culture

Delta CT values from qPCR measurements were tested for normality using Shapiro–Wilk test.

For normally distributed data, differences between treatment groups for each individual gene were evaluated using a one‐way analysis of variance. In the case of nonnormally distributed data, the Friedman test was applied to assess differences between treatment groups.

Densitometry data from Western blot experiments were utilized for statistical tests. Normality of the data was examined using the Shapiro–Wilk test. For each protein marker, a two‐tailed paired t‐test was conducted to identify statistical differences in the protein expression between iron‐treated and BTK‐inhibited samples.

Results of both qPCR and Western blots are depicted as fold changes with error bars denoting the standard error of the mean. Data were not corrected for multiple comparison due to the explorative nature of the cell culture experiments. Statistical significant results were determined with a threshold of p < 0.05.

3. RESULTS

3.1. BTK expression in relation to myeloid cell iron accumulation in post‐mortem human brains

Using IHC on formalin‐fixed paraffin‐embedded (FFPE) brain tissue of controls and MS cases (Table 1), we found the BTK protein to be expressed by neutrophilic granulocytes in the lumen of blood vessels (Figure S1a), in B cells in the leptomeninges (Figure S1c) and perivascular spaces within MS lesions (Figure S1b), and in myeloid cells (i.e., microglia and macrophages) (Figure 1). In normal white matter and normal‐appearing white matter, BTK⁺ myeloid cells were only occasionally found in perivascular spaces of small to medium‐sized blood vessels (Figures 1A,B and S2a–c). In active lesions (Figures 1D and S4a–c) and rims of chronic active lesions (Figures 1C and S3a–c), BTK⁺ myeloid cells were numerous and partially mirrored the distribution of iron‐laden myeloid cells in consecutive sections. Most of these cells either showed an ameboid morphology with processes, indicative of microglial activation, or a foamy cytoplasm without processes, i.e. phagocytosing macrophages. Co‐localization of BTK, CD68, and ferritin light subunit (FTL), a surrogate marker for intracellular iron, was observed at the rim of a chronic active lesion from a representative MS case (Figure 2). By contrast, low numbers of CD68⁺ cells were found throughout inactive lesions, which showed detectable intracellular iron only infrequently. Myeloid BTK expression in inactive lesions was low (Figure 1B), in particular without accumulation of BTK⁺ cells at inactive lesion rims (not shown). Manual cell counting revealed the highest numbers of BTK⁺, TBB⁺, and CD68⁺ cells at the rim of chronic active lesions and in the center of active lesions (Figure 3A–C). BTK⁺ (p < 0.01) and CD68⁺ cells (p < 0.05) were significantly higher at the rim of chronic active lesions compared to the rim of inactive lesions. In addition, the center of active lesions showed significantly higher numbers of BTK⁺ (p < 0.001), CD68⁺ (p < 0.0001), and TBB⁺ cells (p < 0.05) than the center of inactive lesions. Linear regression of BTK⁺ and iron⁺ cell counts across all ROI types was significant (p < 0.001) (Figure 3D). Taken together, this suggests expression of BTK particularly in iron‐laden CD68⁺ microglia and macrophages.

FIGURE 1.

Immunohistochemistry (microglia/macrophage phagocytosis marker CD68, BTK) and iron histochemistry of a representative control white matter (NWM, line A), MS NAWM (line B), a chronic active MS lesion (C) and an active MS lesion (D). Scale bar = 50 μm; Inset scale bar = 20 μm. NAWM, normal‐appearing white matter; NWM, normal white matter.

FIGURE 2.

Representative triple immunofluorescent stainings of chronic active lesion rim from MS patient using DAPI (blue) (A), CD68 (green) (B), FTL (yellow) (C), and BTK (pink) (D). BTK, Bruton's Tyrosine Kinase; CD68, cluster of differentiation of 68; DAPI, 4′,6‐diamidino‐2 phenylindole; FTL, ferritin light chain. Scale bar = 50 μm.

FIGURE 3.

Quantification of BTK (A), TBB (B) and CD68 (C) positive cells in ROIs of controls (n = 10), inactive (n = 10), chronic active (n = 10), and active lesions (n = 7). Dot plot displays the correlation between BTK and TBB, including cell counts of control WM as well as inactive, chronic active, and active MS lesions (D). Each dot represents one patient sample. Colors indicate distinct MS lesion types. NWM, normal white matter; PPWM, periplaque white matter. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

3.2. The inhibition of BTK reduced the proinflammatory response and iron retention in activated HMC3 cells

To assess the effect of BTK inhibition on the inflammatory response and on iron metabolism, we measured mRNA levels in HMC3 cells divided into three treatment groups using real‐time quantitative PCR. Group 1 was cultured under homeostatic conditions in EMEM medium. Group 2 was treated with iron‐dextran in RPMI for iron loading. Group 3 received iron‐dextran, followed by Evobrutinib to inhibit BTK in iron‐loaded cells. All experiments were performed in triplicates. The expression of two microglia markers, 10 selected genes related to proinflammatory microglia activation, and six selected iron‐related genes were analyzed (Table 2). Iron‐dextran loading of cells led to increased mRNA levels of the proinflammatory genes and a decrease of the microglial quiescence marker P2RY12 (Figure 4A). Ferritin light subunit (FTL) mRNA, responsible for intracytoplasmic iron storage, was higher in iron‐treated cells when compared with untreated cells (Figure 4B). Iron loading increased the expression of the ferritin receptor TIM1, the scavenger receptor LRP1, the iron exporter FPN and the FPN antagonist HAMP. BTK inhibition led to a downregulation of genes involved in microglial activation and cytokine release, that is, TLR4 and associated downstream molecules including BTK, NFKB1 (NF‐kB p50), RELA (NF‐kB p65), and NLRP3. mRNA levels of the proinflammatory cytokines IL6 and IL1B were downregulated and P2RY12 was upregulated in Evobrutinib‐treated cells compared with iron‐loaded cells (Figure 4C). BTK inhibition decreased the mRNA levels of TIM1, LRP1, and DMT1, as well as HAMP and FTL. By contrast, FPN was upregulated upon BTK inhibition (Figure 4D). Using immunocytochemistry, we visualized the protein expression of Iba‐1, ferritin, and FPN in iron‐laden (Figure 5A) as well as iron‐laden Evobrutinib‐treated cells (Figure 5B) to confirm the expression of these markers in HMC3 cells. As previously shown with gene expression analysis, quantitative protein level analysis via Western blotting of the investigated iron‐related molecules Dmt‐1, Lrp‐1, Tim‐1, hepcidin, and ferritin (t = 7.309, p = 0.0182; two‐tailed paired t‐test) showed a downregulation upon BTK inhibition. Also, IL‐1b, RelA (t = 4.668, p = 0.0430; two‐tailed paired t‐test) and TLR4 levels showed a decrease in protein expression (Figure 5C).

FIGURE 4.

Fold changes of expression of inflammation‐related (A, C) and iron‐associated genes (B, D) in HMC3 cells. (A, B) Expression in HMC3 cells upon 20 h of iron‐dextran treatment, compared with untreated cells cultured in EMEM medium. (C, D) Each bar depicts fold changes of gene expression in iron‐loaded HMC3 cells with Evobrutinib application, compared with iron‐loaded cells without BTK inhibition. mRNA levels were detected with quantitative real‐time PCR. Error bars represent standard error of the mean (SEM).

FIGURE 5.

Expression of inflammation‐related and iron‐related proteins in HMC3 cells. Representative immunocytochemistry of iron‐loaded only (A) and iron‐loaded Evobrutinib‐treated HMC3 cells (B) stained for Iba‐1, ferritin, and ferroportin (A, B). (C) Differential protein expression of selected markers in iron‐loaded HMC3 cells with BTK inhibition compared with iron‐loaded cells. Samples were normalized to beta‐actin. Each bar depicts the fold changes in protein expression between the two treatment groups, with error bars representing the standard error of the mean (SEM). Statistical significance was determined using two‐tailed paired t‐test on densitometry values and the associated p‐values are indicated by asterisks (*). Asterisks denote the level of significance as follows: * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

In conclusion, BTK inhibition dampened the proinflammatory response as well as iron importers, the iron storage protein ferritin light chain and hepcidin, and increased expression of the homeostatic microglia marker P2ry12 as well as the iron exporter FPN in iron‐laden HMC3 cells. Therefore, BTK inhibition not only reduced the proinflammatory response to iron loading but also the expression of genes implicated in iron uptake and retention in HMC3 cells.

3.3. LRP‐1, TIM‐1, and NLRP3 expression is upregulated at the rim of chronic active MS lesions

To evaluate the expression of proteins of interest in myeloid cells of chronic active lesions, we performed IHC of Lrp‐1 (Figure 6B), Tim‐1 (Figure 6C), and NLRP3 (Figure 6D) on a representative FFPE post‐mortem MS case, harboring a prototypical chronic active iron rim lesion. Qualitatively, these markers were pronouncedly expressed by cells with microglia/macrophage morphology at the rim of chronic active lesions where CD68⁺ myeloid cells were located (Figure 6A). We observed weaker expression of the markers in the lesion center and periplaque white matter of these lesions. Most positive cells exhibited an ameboid morphology and, hence, indicated an activated microglial state, similar to those observed in Figure 1C,D.

FIGURE 6.

Immunohistochemistry of chronic active MS lesion with microglial/macrophage marker CD68 (A), low‐density lipoprotein receptor–related protein 1 (Lrp‐1) (B), the ferritin receptor Tim‐1 (C) and the inflammasome NLRP3. Cells with microglia/macrophage (arrowheads) and astrocyte morphology (arrow) (D). Scale bar = 125 μm, inset scale bar = 25 μm.

4. DISCUSSION

In the current study, we have investigated the expression of BTK on the protein level in a sample of postmortem control and MS brain tissue, and on mRNA and protein level in cultured HMC3 cells. We found BTK to be expressed on neutrophilic granulocytes within blood vessel lumina, B cells in the leptomeninges and perivascular spaces of active MS lesions, and microglia and macrophages throughout active lesion areas and in the rims of chronic active MS lesions. This cellular expression pattern is compatible with previous reports [18, 30, 31, 32, 33]. To the best of our knowledge, this is the first demonstration of BTK expression in myeloid cells in the tissue of postmortem MS brains. We furthermore observed a correlation of myeloid BTK expression and iron accumulation in myeloid cells in actively demyelinating lesion areas, which are active lesion centers and rims of chronic active MS lesions on the histological level. This correlation prompted us to experimentally study the effect of both iron loading and BTK inhibition on microglia in vitro.

For these purposes, we used HMC3 cells as a model. This cell line was made in 1995 via SV40‐dependent immortalization of human microglia [34]. According to previous studies, the cell line expresses microglial markers including Iba‐1 and CD14, but not GFAP, in a resting state. It shows a robust proinflammatory signature upon activation, including upregulation of CD68, MHCII, and CD11b, similar to human primary microglia [34, 35, 36]. We confirmed the expression of CD68 and Iba‐1 of our cultured HMC3 cells on mRNA level and protein level. This cell line has been used in numerous studies with various aims including modeling of oxidative stress response [37, 38] or proinflammatory responses upon toxin stimulation [39] or bacterial exposure [40]. Moreover, iron‐overload experiments with the application of ferric ammonium citrate (FAC) have already been performed using HMC3 cells [41]. Upon FAC iron loading, mRNA expression levels of both the divalent ion importer channel (DMT1) and the iron exporter FPN increased, which we were able to replicate in our current study with iron‐dextran loading. Moreover, the mRNAs of the cytokines IL1B, IL6, and TNF were shown to be upregulated [34], which we confirmed here for IL1B and IL6 on the mRNA level. It is still a matter of debate whether the iron loading of microglia and macrophages in the rims of chronic active lesions is the cause or consequence, or both at the same time, of their proinflammatory activation. Recent in vitro work by the group of Pitt et al. found increased iron uptake in “M1”‐prepolarized iPSC‐derived microglia [42], while they reported no downstream effect of iron loading on their ROS or cytokine production. However, the authors discuss the probable dependency of the experimental results on the iron compound used for loading experiments. The findings of Carota et al. and of our study provide some argument for the proinflammatory effect of iron loading on microglia, at least under in vitro conditions, which is in line with earlier in vitro work [26, 43]. A previous study has demonstrated the activation of heme scavenging upon TLR7 activation via activation of BTK and subsequent induction of the heme‐hemopexin complex receptor Lrp1 expression [21]. That study, therefore, linked Lrp‐1‐mediated uptake of iron‐containing heme‐hemopexin complexes to BTK functioning in microglia. Other results from our work implicate hemoglobin‐derived iron loading of microglia and macrophages at the rim of chronic active lesions via the haptoglobin‐hemoglobin complex receptor CD163. Here, we have accordingly shown that also Lrp‐1 is both expressed in microglia and macrophages at iron rims of chronic active MS lesions of human postmortem brain tissue and downregulated in microglia in vitro upon BTK inhibition using Evobrutinib. In addition, we show the downregulation of the ferritin receptor Tim‐1 on microglia upon BTK inhibition. By contrast, we found the iron exporter FPN to be upregulated upon application of Evobrutinib, which in concert suggests a reduction of microglial iron load. It remains to be clarified whether such an inhibition of iron import and stimulation of iron export is beneficial for local tissue integrity upon BTK inhibition in patients. Taken together, we thus show both a dampening of a proinflammatory response in microglia and a decrease of iron import/retention as well as an increase in iron export upon BTK inhibition in vitro. Our study has important implications for possible outcomes of currently ongoing clinical trials using BTK inhibitors. As of May 2023, there are seven clinical trials listed (http://clinicaltrials.gov; search terms: multiple sclerosis, Bruton's tyrosine kinase). One of these trials (NCT04742400), testing tolebrutinib in MS patients, reportedly uses PRL development at 48 weeks as primary outcome parameter. Our data suggest that PRs, which are formed by iron‐laden microglia at the edge of chronic active lesions, might indeed be attenuated by BTK inhibition in MS patients. Furthermore, our data imply that this attenuation of rims is accompanied by a dampening of proinflammatory activation of microglia at the edges of chronic active lesions, rendering the PR sign a potentially valid tool of monitoring chronic inflammation also under therapeutic conditions. One limitation of our study is the focused approach on the effect of Evobrutinib on the HMC3 microglial cell line with regard to a selected set of iron‐related proteins. However, we believe we are able to link the Evobrutinib effect on microglia with the regulation of iron‐related proteins expressed in microglia in the MS tissue. Therefore, we are confident that our study provides the basis for the use of the PR sign in clinical trials on BTK inhibitors and, possibly, in the assessment of individual people with MS treated with BTK inhibitors.

AUTHOR CONTRIBUTIONS

I.W., S.H. and A.H. designed the study. A.S., C.R. and G.T. acquired the data. T.B. and H.L. helped with the interpretation of the data. H.L., J.B. and U.K. provided patient samples from the archives of the Center for Brain Research, Medical University of Vienna. R.H. and S.H. provided patient samples of the archives of the Department of Neurology, Medical University of Vienna. T.K. gave methodological advice in terms of gene expression analysis. The manuscript was written by S.H., T.B. and A.H. and reviewed by all authors.

Supporting information

Data S1. Supporting information.

Steinmaurer A, Riedl C, König T, Testa G, Köck U, Bauer J, et al. The relation between BTK expression and iron accumulation of myeloid cells in multiple sclerosis. Brain Pathology. 2024;34(5):e13240. 10.1111/bpa.13240

DATA AVAILABILITY STATEMENT

Data are available based upon reasonable request.