BSc Biomedical Sciences
Department of Experimental Immunology
Research Project
Folate receptor-beta expression reflects macrophage density in distinct
synovial pathotypes of rheumatoid arthritis
Jeffrey van der Krogt 10543481 April 2020
Assessor: Examiner:
1 ABSTRACT
Background: Personalizing RA treatment by means of patient stratification might reduce the
delay in effective rheumatoid arthritis (RA) treatment. One way to stratify RA patients is by means of synovial tissue pathotyping. Recently, the exploration of synovial tissue biomarkers has led to the proposition of three distinct synovial pathotypes: lympho-myeloid, diffuse myeloid, and pauci-immune/fibroid. Initial results have demonstrated RA synovial pathotype stratification to potentially predict RA treatment response. Over the past decade, positron emission tomography (PET) imaging has emerged as a non-invasive alternative of synovial biopsy for synovial tissue analysis. However, it is unknown if PET imaging has the potential to contribute to patient stratification. One of the parameters analyzed during patient
stratification based on synovial pathotype is macrophage infiltration. The folate receptor-beta (FR-β) proved to correlate to macrophage count and might therefore be a useful PET target during synovial pathotype stratification. The aim of this study is to examine FR-β expression in the three distinct RA synovial pathotypes. In addition, FR-β expression in relation to macrophage polarization was examined.
Methods: RA patients underwent arthroscopic biopsy of an actively inflamed knee or ankle.
Biopsy sections were immunohistochemically stained in order to stratify each patient into one synovial pathotype group. Next, biopsy sections were immunofluorescently stained in order to determine FR-β expression and macrophage polarity. Relationships between FR-β expression and macrophage count/polarity were determined for each patient sample.
Results: FR-β expression was significantly lower in the fibroid/pauci-immune synovial tissue
pathotype as compared to the diffuse myeloid and lympho-myeloid synovial pathotypes. Within each pathotype group, a significant positive correlation between FR-β and CD68 expressions was found. Despite a significantly decreased FR-β expression among
CD163+CD68+ (M2) macrophages, macrophage polarization did not affect the correlation between FR-β expression and macrophage count significantly.
Conclusions: The results of this study highlight the potential to quantify the macrophage
count within distinct synovial pathotypes based on FR-β expression. Therefore, FR-β could be used as a target for RA patient stratification by means of PET imaging of the inflamed
synovium. Further research should explore the ability to perform synovial pathotype stratification using a combination of two PET tracers.
2 INTRODUCTION
Rheumatoid arthritis (RA) is a chronic autoimmune disease with a worldwide prevalence of approximately 5-10 patients per 1000 adults [1]. This disease is primarily characterized by inflammation of the synovial tissue in the joints, thereby damaging cartilage and bone tissue [2]. Although the development and application of biological disease modifying anti-rheumatic drugs (bDMARDs) has dramatically improved the prognosis of patients affected by RA, delay in effective treatment remains a major challenge [3]. Subjecting patients to ineffective
treatments allows the expansion of structural damage and leads to poor long-term outcomes, indicating the need to improve RA treatment efficacy. One way to improve treatment efficacy is by applying personalized RA treatment through patient stratification, which requires the identification of appropriate biomarkers [4, 5]. While peripheral blood biomarkers, such as rheumatoid factor and anti-citrullinated peptide antibodies (ACPA) proved relatively high specificity and the ability to predict RA development [6], these markers are expressed in only 70-80% of RA patients. With the synovium as a primary target of inflammation in RA, detailed analysis of the inflamed synovial tissue has emerged as a promising alternative approach in the search of RA biomarkers for patient stratification [7, 8].
Exploring the use of synovial tissue biomarkers in personalizing RA treatment has led to the proposition of three distinct synovial pathotypes with a specific histology and gene expression signature: lympho-myeloid, diffuse myeloid, and pauci-immune/fibroid [9]. Whereas the lympho-myeloid pathotype is characterized by B cell aggregates and a high occupation of plasma cells, macrophages dominate in the diffuse myeloid pathotype. The pauci-immune/fibroid pathotype consists of a low overall immune cell count. Initial results demonstrated cellular heterogeneity in RA synovial tissue to impact clinical outcome to therapy [10]. Whereas the myeloid pathotype exhibited good response to treatment with anti-tumor necrosis factor (TNF)-α, a significantly lower response to TNF-α was achieved in patients with a predominant fibroid/pauci-immune synovial pathotype. This finding underscores the potential contribution of patient stratification to the improvement of RA treatment efficacy.
In order to determine the synovial pathotype based on histopathological analysis, the synovial tissue must be obtained for which ultrasound-guided and arthroscopy-assisted synovial biopsy are widely-used techniques. Although these techniques enable excellent histological evaluation of the synovial tissue, this technique is rather invasive and allows
3
analysis of only one or a limited amount of joints. In response, positron emission tomography (PET) has emerged as a non-invasive and widely reaching alternative for the evaluation of synovial inflammation in RA patients [11]. Through the use of target-specific tracers, PET imaging may reveal information on tissue composition with a high sensitivity and specificity, potentially also making it a suitable technique for synovial pathotype stratification [12].
One of the parameters analyzed during synovial pathotype stratification is macrophage infiltration. Macrophages are critically involved in the pathogenesis of RA with the
production of inflammatory cytokines and the contribution to destruction of cartilage and bone tissue through multiple mechanisms. In addition, synovial macrophage count has been highlighted as a sensitive biomarker for response to treatment in patients with RA during previous research [13]. Recently, the PET tracer [18F]fluoro-PEG-folate specifically targeting the folate receptor-β (FR-β) was introduced [14]. This receptor is strongly expressed on activated rather than quiescent macrophages, both in the lining and in the sublining of synovial tissue from RA patients [15]. While a significant positive correlation between FR-β expression and the number of macrophages in human RA synovial tissue has been described in a previous study [16], the possibility of quantifying macrophages through FR-β expression in the context of synovial pathotype stratification has not been explored yet. Therefore, the aim of this study is to examine the relation between FR-β expression and macrophage count with regard to the three distinct synovial pathotypes in RA patients. Beside macrophage count, the role of macrophage heterogeneity in RA disease activity increasingly gains attention. One aspect of macrophage heterogeneity is covered by macrophage M1-M2
polarization balance. In vitro studies revealed that granulocyte macrophage colony stimulating factors (GM-CSF)-derived M1 macrophages are pro-inflammatory and initiate Th1 responses, whereas M-CSF-derived M2 macrophage are largely characterized by anti-inflammatory and tissue repair activity [17]. Since discrepant results regarding the influence of macrophage polarity on FR-β expression have been described in literature [18, 19], the relation between FR-β expression and macrophage polarity will also be further explored in the current study.
MATERIAL AND METHODS Patients
RA patients with at least one inflamed ankle or knee who underwent arthroscopy for
4
Clinical Immunology at the Amsterdam University Medical Center, location AMC. Clinical inflammation was defined as the occurrence of both joint pain and swelling at physical examination. RA diagnosis was confirmed following the 1987 American College of
Rheumatology (ACR) criteria [20]. Upon inclusion, the following patient demographics were collected: age; gender; disease duration; rheumatological medication; disease activity score in 28 joints (DAS28); visual analogue score (VAS)-patient; IgM-Rheumatoid factor (RF) levels using IgM-RF ELISA (Sanquin, Amsterdam, the Netherlands until 2009 and thereafter Hycor Biomedical, Indianapolis, IN); anti-cyclic citrullin peptide antibody (ACPA) using CCP2 ELISA CCPlus (Eurodiagnostica, Nijmegen, the Netherlands); erythrocyte sedimentation rate (ESR); and serum levels of C-reactive protein (CRP). All patients gave informed consent. The study was approved by the Amsterdam University Medical Center Medical Ethics Committee and performed according to the Declaration of Helsinki.
Synovial biopsies
Arthroscopy-guided synovial biopsies were performed in a clean procedure room. All
biopsies were taken from an inflamed ankle or knee after injection of local anesthesia, using a small-bore arthroscope (Storz, Tuttlingen, Germany) together with two portals for
macroscopic examination. A minimum of 12 tissue specimens were obtained from each patient. Directly after biopsy, synovial tissue was snap-frozen with liquid nitrogen and stored at -80°C, followed by cryosectioning at -20°C. 5μm thick sections were anonymously stored at -80°C within the departmental biobank.
Tissue quality assessment
Hematoxylin and eosin (H&E) staining was performed on one synovial tissue biopsy section of each patient included in order to examine biopsy size and morphology. The applied H&E staining protocol is attached in Supplementary file 1. In brief, sections were fixed with 50 μL 4% paraformaldehyde (PFA), followed by staining with 150 μL hematoxylin solution for 8 minutes and eosin solution for 3 minutes respectively. Sections were dehydrated and mounted with entellan mounting medium. After H&E staining, biopsy size was assessed by one reader in order to assure enough tissue for subsequent immunohistological staining. Patients with sections too small for histological analysis were excluded from the study. Following, tissue morphology was scored semi-quantitatively by two readers, based on the following
assessment: (0= tissue damage and overlay; 1= tissue damage or overlay; 2= no tissue damage or overlay). Sections with a quality score of 0 were considered for exclusion.
5 Synovial tissue pathotype classification
After synovial tissue biopsy quality assessment, immunohistochemical (IHC) staining was applied to four sections of each patient using 3,3’-diaminobenzidine (DAB) and primary antibodies targeting B cells (CD20), T cells (CD3), macrophages (CD68) and plasma cells (CD138). The complete IHC staining protocol is attached in Supplementary file 2. In brief, sections were firstly fixed with acetone for 10 minutes at room temperature (RT) and blocked with 10% casein for 10 minutes at RT. Subsequently, each of four sections of every patient was incubated for 60 minutes at RT with 70 μL with one of the following primary antibodies respectively: (1) CD3 mouse anti-human, IgG1k, 500μg/ml (Vector clone VP-C429), 1:30; (2) CD20 mouse human, IgG2a, 126μg/ml (Dako clone L26), 1:25; (3) CD68 mouse anti-human, IgG1k, 237μg/ml (Dako clone M0718), 1:800; (4) CD138 mouse anti-anti-human, IgG1k, 500μg/ml (Biolegend clone MI15), 1:110. After incubation of the primary antibody, 80μL of polyclonal goat-anti mouse – HRP 1,0mg/ml (Dako clone P0447), 1:400 was added for 30 minutes at RT. Immune cell infiltration was visualized by the application of DAB staining to each section. Degrees of immune cell infiltration were scored semi-quantitatively (0-4 scale) by two readers. Based on IHC scores for each specific immune cell type, synovial tissue biopsies were classified into one of three classes following the classification criteria as
introduced by Humby et al. 2019 [21]: fibroid/pauci-immune (F) CD68SL≤1 and no presence of CD3, CD20, and CD138; diffuse-myeloid (D) CD68SL≥2, CD20≤1 and/or CD3≥1,
CD138≤2; lymphoid-myeloid (L) presence of grades 2-3 CD20+ aggregates, CD20≥2 and/or CD138≥2.
IF staining and imaging of synovial biomarkers
Triple immunofluorescence (IF) staining was performed in order to analyze the expression of FR-β, CD68 (macrophages), and CD163 (M2 macrophages). The complete IF staining protocol is attached in Supplementary file 3. In brief, each section was incubated for 60 minutes with 80 μL of all of the following three primary antibodies consecutively: (1) FR-β rat human, IgG2a (Matsuyma batch March 2016, G. Jansen), 1:40; (2) CD163 rabbit anti-human, IgG, 700μg/ml (Abcam clone EPR19518), 1:100; (3) CD68 mouse anti-anti-human, IgG1k, 237μg/ml (Dako clone M0718), 1:800. Thereafter, 80 μL of a secondary antibody cocktail was added for 30 minutes at RT consisting of the following three secondary antibodies: (1) 1/3 donkey anti-rat Alexa Fluor 594, IgG, 2mg/ml (ThermoFisher cat. 21209), 1:70; (2) 1/3 donkey anti-rabbit Alexa Fluor 488, IgG, 2mg/ml (ThermoFisher cat. A-21206), 1:70; (3) 1/3 donkey anti-mouse Alexa Fluor 647, IgG, 2mg/ml (ThermoFisher cat.
6
A-31571), 1:70. After IF staining, images of four section parts (most intense; least intense; most representative; lining) were obtained with the Leica SP8-X confocal microscope (Leica Microsystems, Wetzlar, Germany), using a 20x immersion oil objective with zoom factor 1.34. Laser intensity settings were determined based on signal intensities of both a positive and negative control section.
Quantification of IF expression and statistical analysis
ImageJ (U.S. National Institutes of Health, Bethesda, Maryland, USA [22]) was used to quantify immunofluorescent expression of FR-β, CD68, and CD163 per section. First, in all four images acquired from each section, expression intensity was normalized based on DAPI expression intensity. Following, integrated density of FR-β, CD68, and CD163 was calculated as the product of biomarker expression area multiplied by biomarker expression intensity. Averages of all four integrated densities from each section were calculated and used for further statistical analysis. Consecutively, the following features were analyzed for each section: (1) mean FR-β integrated density; (2) mean CD68 integrated density to determine macrophage count; (3) mean FR-β integrated density within CD68+/CD163+ and
CD68+/CD163- areas separately to distinguish between M2 and non-M2 macrophages
respectively; and (4) mean CD163/CD68 integrated density division to determine presence of M2 macrophages within total pool of macrophages.
Statistical analyses
Descriptive statistics were used for generation of patient characteristics. Square-root
transformation was used to improve distribution normality within synovial pathotype groups. In order to determine the distribution of results within each pathotype group, the Shapiro-Wilk test was performed to test for normality and the Levene statistic was applied to test for homogeneity of variance. Difference in integrated density between groups were analyzed using the one-way Anova test with Tukey post-hoc test (normal distribution and equal variances fulfilled) or the Kruskal-Wallis test (normal distribution and equal variances not fulfilled). A paired sample t-test was used to compare the expression of two different IF markers within one section. The relation between FR-β and CD68 integrated densities was examined using the Pearson correlation. A bar graph, box plots and a scatter plot were used for graphical visualization of statistical analyses, generated through the statistical program SPSS (IBM Corp. Released 2017. IBM SPSS Statistics for Windows, Version 26.0. Armonk, NY).
7 RESULTS
Identification of distinct synovial pathotypes within patients
In total, thirty-one active RA patients were included in this study. Two patients were excluded after initial tissue section examination due low tissue quality (n=1) or too small biopsy section size for further examination (n=1). Of the remaining twenty-nine patients, twenty had
synovial tissue biopsy sections of good quality and nine had sections of sufficient quality. Examples of H&E stained synovial tissue biopsy sections with quality scores 0-2 are displayed in figure 1.
Score 0: insufficient quality Score 1: sufficient quality Score 2: good quality
Figure 1. Light microscopic images of H&E stained sections from synovial tissue biopsies classified by semi-quantitative quality score.
Among patients included in the current study, ten patients portrayed a fibroid/pauci-immune pathotype, nine a diffuse myeloid pathotype, and ten a lympho-myeloid pathotype. Semi-quantitative scores of separate targeted immune cell biomarkers are illustrated for each distinct pathotype in figure 2. In addition, representative light microscopy images for each separate pathotype are displayed in figure 3.
Figure 2. Clustered bar chart of semi-quantitative IHC score medians for each synovial tissue immune cell biomarker, categorized based on synovial pathotype.
8 CD138 CD20 CD3 CD68 H&E Fibroid/Pauci-immune Diffuse-myeloid Lympho-myeloid
Figure 3. Light microscopic images of IHC stained sections from synovial tissue biopsies classified by pathotype based on the following semi-quantitative assessment: fibroid/pauci-immune CD68SL≤1 and no presence of CD3, CD20, and CD138; diffuse-myeloid CD68SL≥2, CD20≤1 and/or CD3≥1, CD138≤2; lymphoid-myeloid presence of grades 2-3 CD20+ aggregates, CD20≥2 and/or CD138≥2.
In line with previous literature [21], the fibroid/pauci-immune group in the current study portrays a significantly decreased C-reactive protein (CRP) level as compared to the lympho-myeloid group (median=3,00 vs. median=23,95; p=0,034). Patient characteristics within each distinct pathotype group are set out in table 1.
Table 1. Patient characteristics within each of three distinct synovial pathotypes. Patient characteristic
Fibroid/pauci-immune (n=10) Diffuse myeloid (n=9) Lympho-myeloid (n=9) Age, mean (SD) 54,50 (15,2) 54,17 (5,8) 51,38 (8,1)
Disease duration in days, median (SD) 328 (1988) 602 (765) 2600 (2247) No. (%) of rheumatoid factor positive 5 (71) 3 (38) 8 (80)
No. (%) of ACPA positive 2 (40) 3 (43) 7 (78)
DAS28, mean (SD) 5,1 (1,1) 5,7 (1,2) 5,5 (1,1)
VAS patient, mean (SD) 53,8 (20,3) 43,8 (29,3) 61,3 (19,7)
ESR, mean (SD) 26,0 (21,0) 42,3 (34,0) 47,8 (41,5)
CRP, median* 4,00 39,00 23,95
SD: standard deviation; No.: number; DAS28: Disease Activity Score-28 joints; VAS: Visual Analogue Score; ESR: Erythrocyte Sedimentation Rate; CRP: C-Reactive Protein; *p<0,05; **p<0,01.
Low FR-β expression by fibroid/pauci-immune correlates with low macrophage density in distinct synovial pathotypes
Of twenty-nine RA patients, confocal images with immunofluorescent expressions of FR-β, CD68, and CD163 were obtained from one section. A representative example for the triple IF
9
staining is shown in figure 3. For each biomarker expressed, integrated densities were calculated as the product of expression area multiplied by the expression intensity.
H&E DAPI FR-β
CD68 CD163 Merged
Figure 3. Confocal images of a triple IF stained section from a synovial tissue biopsy with primary antibodies targeting FR-β, CD68, and CD163. Magnification 20x; Zoom factor 1.34.
FR-β integrated densities of synovial tissue biopsy sections within each pathotype group fulfilled the assumptions of equal distribution and homogeneous variances. A significantly lower mean FR-β integrated density was found within the fibroid/pauci-immune group as compared to the diffuse myeloid and lympho-myeloid groups (mean(SD)=1,81(0,39) vs. mean(SD)=3,74(0,51) vs. mean(SD)=3,89(0,48); p=0,005) (Fig. 4).
Figure 4. Significantly lower FR-β integrated density within fibroid/pauci-immune pathotype as compared to the diffuse myeloid and lympho-myeloid pathotypes. *p<0.05; **p<0.01, one-way Anova test with Tukey post-hoc test; SqRt, square root.
** *
10
CD68 Integrated Density (SqRt transformed)
FR -β In teg rat ed D ens it y (Sq R t t rans for m ed )
Previous research has illustrated the fibroid/pauci-immune pathotype to contain a low macrophage cell count as compared to the diffuse myeloid and lympho-myeloid pathotypes [9]. Furthermore, a correlation between the expressions of FR-β and CD68 in synovial tissue of RA patients has been described in literature [16]. Based on these findings, the significantly lower FR-β expression in the fibroid/pauci-immune pathotype was expected. Furthermore, macrophage count is suspected to significantly contribute to this difference in FR-β
expression between synovial pathotypes. In order to examine the contribution of macrophage count, the correlation between macrophage count and FR-β expression is determined in each synovial pathotype separately first. During this study, the correlation between FR-β and CD68 integrated densities appeared to be significant within each distinct synovial pathotype group (fibroid/pauci-immune (F): correlation (r)=0.718, p=0.019, diffuse myeloid (D): r=0.750,
p=0.020, lympho-myeloid (L): r=0.869, p=0.001) (Fig. 5).
Figure 5. Significant positive correlation between FR-β and CD68 integrated densities with a distinction between three known synovial tissue pathotypes. SqRt, square root.
These results, together with the findings of previous studies, brought up the assumption that a low macrophage count mainly accounts for the low FR-β expression within the fibroid/pauci-immune pathotype group. In order to prove this assumption, an analysis of covariance (ANCOVA) was performed. Homogeneity of regression slopes, needed for further analysis using ANCOVA, was met among all distinct synovial pathotypes (F=0.433; p=0.654).
11
ANCOVA revealed a non-significant difference in FR-β integrated density between synovial tissue pathotypes after correction for macrophage count based on CD68 integrated density (F: mean(SE)=2.90(0,33) vs. D: mean(SE)=3.25(0.31) vs. L: mean(SE)=3.24(0.30); p=0.737), confirming the assumption significant contribution of macrophage count to FR-β expression.
Total macrophage count not affected by low FR-β expression on M2 macrophages
Beside macrophage count, macrophage polarity proved to affect inflammatory activity in the synovium of RA patients. Whereas granulocyte macrophage colony stimulating factors (GM-CSF)-derived M1 macrophages are pro-inflammatory and initiate Th1 responses, M-CSF-derived M2 macrophage are characterized by anti-inflammatory activity [17]. During this study, for each patient the FR-β signal expressed by CD163-CD68+ (non-M2) and
CD163+CD68+ (M2) macrophages was determined respectively. Comparison of the FR-β signal expressed by CD163- and CD163+ macrophages using a paired sample t-test generated a significantly higher FR-β expression by CD163-CD68+ (non-M2) macrophages as compared to CD163+CD68+ (M2) macrophages (mean(SD)=0.520(0.259) vs. mean(SD)=0.359(0.188);
p=0.002) (Fig. 6).
Figure 6. CD163+ macrophages express significantly less FR-β signal compared to CD163 -macrophages. **p<0.01, paired sample t-test.
In order to examine the effect of this difference in FR-β expression on the ability to reliably quantify macrophage count within distinct synovial pathotypes based on FR-β expression, the proportion of M2 macrophages within the total group of CD68+ macrophages was determined
12
in each synovial pathotype. Based on the Kruskal-Wallis test, no significant difference between the proportion of M2 macrophages was found between the three distinct synovial pathotypes (F: median=0.171 vs. D: median=0.246 vs. L: median=0.268; p=0.002) (Fig. 7).
Figure 7. Proportion of CD163+ macrophages among overall CD68+ macrophage population does not differ significantly between distinct synovial pathotypes.
These results suggest macrophage polarity to have only minimal influence on the significant correlation between macrophage count and FR-β expression. To confirm this hypothesis, a multiple regression analysis was performed on each pathotype group separately. This test proved the proportion of M2 macrophages to not significantly influence the FR-β expression (F: p=0.862, D: p=0.479, L: p=0.372), once again underscoring the ability to determine macrophage density in synovial tissue based on FR-β expression.
DISCUSSION
This is the first study to evaluate the ability of the macrophage specific FR-β as a potential PET target to quantify the number of macrophages in distinct synovial pathotypes of RA. During this study, the fibroid/pauci-immune synovial tissue pathotype demonstrated to express a significantly lower FR-β signal as compared to the diffuse myeloid and lympho-myeloid pathotypes. In addition, a significant positive correlation between FR-β and CD68 integrated densities has been confirmed, suggesting the low FR-β expression in the
fibroid/pauci-immune synovial pathotype to originate from a low macrophage count. This positive correlation between FR-β and CD68 integrated densities has been extrapolated to each distinct synovial pathotype. In response to discrepancy in existing literature, this study
13
shows a significantly decreased FR-β integrated density on CD163+ (M2) macrophages. Still, unequal proportions of CD163+ cells within the region of CD68+ cells between distinct synovial pathotype proved not to affect the FR-β integrated density significantly. Together, these findings underscore the ability of FR-β to quantify macrophage count in distinct
synovial tissue pathotypes of RA patients, highlighting the potential of this receptor as a target for synovial pathotype stratification based on PET.
Previous research has illustrated the fibroid/pauci-immune pathotype to contain an overall low immune cell count [9]. Despite the knowledge of FR-β to function as a folate transporter complex mainly on activated macrophages, the expression of this receptor in distinct synovial pathotypes was not further examined yet. This study describes a significantly lower FR-β expression in synovial tissue with a dominating fibroid/pauci-immune pathotype, correlating significantly with the expression of CD68. This correlation is in concordant to earlier findings of FR-β to correlate with macrophage count in synovial tissue [16]. Additionally, this study extrapolates the correlation between FR-β expression and macrophage count within synovial tissues of distinct synovial pathotypes. One way the correlation between FR-β expression and macrophage count could have been divergent in synovial tissue with a predominating fibroid/pauci-immune pathotype, is through aspecific binding of the FR-β binding primary antibodies during IF staining. During confocal imaging, aspecific binding of this antibody to collagen tissue was observed (Supplementary file 4). Analysis of fibroid synovial tissue biopsy composition by polymerase chain reaction must confirm this observation to be resulting from aspecific binding. However, since only minimal FR-β signal from aspecific binding has been included in the analysis of the current study due to appropriate thresholding of the IF signal, aspecific binding is expected to not affect the correlation between FR-β and macrophage count. A second theory by which the correlation between FR-β expression and macrophage count could have been affected was through a different FR-β expression between macrophages of a distinct polarity. Due to a discrepancy in the FR-β signal expressed by M1 (pro-inflammatory) and M2
(anti-inflammatory/homeostatic) macrophages in existing literature [18, 19], this possibility was explored during the current study. Indeed, a lower FR-β expression was found for M2 macrophages, but this did not affect the relation between FR-β expression and macrophage count since the proportion of M2 macrophages did not differ significantly between synovial pathotype groups. In addition, a lower FR-β expression by M2 macrophages might be beneficial for the utility of FR-β as a PET-target during synovial pathotype stratification,
14
since M2 macrophages are thought to have a anti-inflammatory function and would thereby correlate negatively with RA disease activity [23]. At last, it is important to keep in mind that the rigid M1/M2 classification is questionable, and further research with bigger cohorts has to be performed to investigate the FR-β expression in relation to macrophage polarity.
This study encountered several limitations. First, the limited amount of patients
included in each synovial pathotype group, together with the expression analysis of one tissue section only, resulted in a decreased power of the results obtained. Second, during confocal imaging of the synovial tissue section, areas were selected based on the intensity of the FR-β signal expressed. Since the researcher was not blinded for synovial pathotype, areas in favor of the study’s results might have been selected. In order to extrapolate the results of the current study to the entire RA population, a higher power has to be achieved through the examination of larger cohorts and a random selection of areas to be imaged during IF expression analysis. The latter issue can be realized by randomly selecting five fragments of DAPI positive areas, instead of selection based on FR-β expression only.
Considering the classification criteria of synovial pathotype stratification [21],
multiple immune cell types need to be quantified in order to reliably classify a synovial tissue biopsy. Although an extremely low FR-β expression may indicate a predominating
fibroid/pauci-immune pathotype, reliable pathotype stratification by means of PET imaging requires simultaneous analysis of multiple targets. The main distinction between diffuse myeloid and myeloid pathotypes lies in a higher B cell count within the lympho-myeloid predominating synovial tissue. Therefore, B cell targeting in combination with macrophage targeting could be advantageous for synovial tissue pathotype stratification. Signal intensity from zirconium-89 labeled rituximab has proven to be associated with 24 weeks treatment response. However, quantifying immune cells by labeling a drug does not improve treatment efficacy as treatment has already started. Another possible target for combination with an FR-β targeting PET tracer is the T cell. Targeting the CD3 part of T cell receptor with a radiolabeled Fab fragment has the advantage of not depleting T cells, and as such does not impact on the pathogenesis or disease course of RA. Parallel to the
development of a second immune specific PET tracer, the construct validity of the findings of the current study can be examined during a prospective study that compares FR-β expression by PET imaging with results from a synovial biopsy. In order to ensure a high amount of inclusions with a heterogeneous pool of RA patients, a prospective study will be performed. If
15
the quantification of specific immune cell types within the synovial tissue of active RA patients based on PET matches the results obtained by synovial tissue biopsy analysis, this imaging technique could serve as a non-invasive alternative for synovial tissue pathotype stratification.
REFERENCES
1. Cross, M., et al., The global burden of rheumatoid arthritis: estimates from the global burden of disease 2010 study. Ann Rheum Dis, 2014. 73(7): p. 1316-22.
2. Scott, D.L., F. Wolfe, and T.W. Huizinga, Rheumatoid arthritis. Lancet, 2010. 376(9746): p. 1094-108.
3. Romao, V.C. and J.E. Fonseca, Major Challenges in Rheumatology: Will We Ever Treat Smarter, Instead of Just Harder? Front Med (Lausanne), 2019. 6: p. 144.
4. Romao, V.C., et al., Right drug, right patient, right time: aspiration or future promise for biologics in rheumatoid arthritis? Arthritis Res Ther, 2017. 19(1): p. 239.
5. Cuppen, B.V., et al., Personalized biological treatment for rheumatoid arthritis: a systematic review with a focus on clinical applicability. Rheumatology (Oxford), 2016. 55(5): p. 826-39. 6. Hensvold, A.H., et al., How well do ACPA discriminate and predict RA in the general
population: a study based on 12 590 population-representative Swedish twins. Ann Rheum Dis, 2017. 76(1): p. 119-125.
7. Orr, C., et al., Synovial tissue research: a state-of-the-art review. Nat Rev Rheumatol, 2017.
13(10): p. 630.
8. Humby, F.C., et al., Can Synovial Pathobiology Integrate with Current Clinical and Imaging Prediction Models to Achieve Personalized Health Care in Rheumatoid Arthritis? Front Med (Lausanne), 2017. 4: p. 41.
9. Pitzalis, C., S. Kelly, and F. Humby, New learnings on the pathophysiology of RA from synovial biopsies. Curr Opin Rheumatol, 2013. 25(3): p. 334-44.
10. Dennis, G., Jr., et al., Synovial phenotypes in rheumatoid arthritis correlate with response to biologic therapeutics. Arthritis Res Ther, 2014. 16(2): p. R90.
11. Bruijnen, S.T., et al., Present role of positron emission tomography in the diagnosis and monitoring of peripheral inflammatory arthritis: a systematic review. Arthritis Care Res (Hoboken), 2014. 66(1): p. 120-30.
12. Jones, T. and D. Townsend, History and future technical innovation in positron emission tomography. J Med Imaging (Bellingham), 2017. 4(1): p. 011013.
13. Haringman, J.J., et al., Synovial tissue macrophages: a sensitive biomarker for response to treatment in patients with rheumatoid arthritis. Ann Rheum Dis, 2005. 64(6): p. 834-8. 14. Chandrupatla, D., et al., The folate receptor beta as a macrophage-mediated imaging and
therapeutic target in rheumatoid arthritis. Drug Deliv Transl Res, 2019. 9(1): p. 366-378. 15. Xia, W., et al., A functional folate receptor is induced during macrophage activation and can
be used to target drugs to activated macrophages. Blood, 2009. 113(2): p. 438-46.
16. van der Heijden, J.W., et al., Folate receptor beta as a potential delivery route for novel folate antagonists to macrophages in the synovial tissue of rheumatoid arthritis patients. Arthritis Rheum, 2009. 60(1): p. 12-21.
17. Verreck, F.A., et al., Human IL-23-producing type 1 macrophages promote but IL-10-producing type 2 macrophages subvert immunity to (myco)bacteria. Proc Natl Acad Sci U S A, 2004. 101(13): p. 4560-5.
18. Tsuneyoshi, Y., et al., Functional folate receptor beta-expressing macrophages in osteoarthritis synovium and their M1/M2 expression profiles. Scand J Rheumatol, 2012.
16 19. Shen, J., et al., Folate receptor-beta constitutes a marker for human proinflammatory
monocytes. J Leukoc Biol, 2014. 96(4): p. 563-70.
20. Arnett, F.C., et al., The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum, 1988. 31(3): p. 315-24.
21. Humby, F., et al., Synovial cellular and molecular signatures stratify clinical response to csDMARD therapy and predict radiographic progression in early rheumatoid arthritis patients. Ann Rheum Dis, 2019. 78(6): p. 761-772.
22. Schneider, C.A., W.S. Rasband, and K.W. Eliceiri, NIH Image to ImageJ: 25 years of image analysis. Nat Methods, 2012. 9(7): p. 671-5.
23. Tardito, S., et al., Macrophage M1/M2 polarization and rheumatoid arthritis: A systematic review. Autoimmun Rev, 2019. 18(11): p. 102397.
SUPPLEMENTARY FILES
Supplementary file 1. The haematoxylin and eosin staining protocol.
Supplementary file 2. The immunohistochemical staining protocol.
Supplementary file 3. The immunofluorescence staining protocol.
