Absence of osteoclastogenic response to pooled polyclonal anti-citrullinated peptide/protein antibodies from patients with rheumatoid arthritis in vitro

Abstract

Anti-citrullinated peptide/protein antibodies (ACPAs) are specific serological markers for rheumatoid arthritis (RA), but their direct role in promoting osteoclastogenesis remains controversial. This study aims to determine whether polyclonal ACPAs purified from RA patients can directly drive osteoclastogenesis in vitro. Polyclonal ACPAs were purified from pooled sera of 37 ACPA positive RA patients using Protein G and multi-citrullinated-epitope peptide (MCEP) conjugated affinity chromatography, with control IgG from 10 ACPA negative RA patients and a healthy donor. The purity and specific binding capability to citrullinated vimentin were verified by Coomassie Blue staining and western blot. CD14 + monocytes from healthy donors and RA patients were induced to differentiate into osteoclasts with M-CSF and RANKL, in the presence or absence of ACPAs. Osteoclast formation was evaluated by tartrate resistant acid phosphatase (TRAP) staining, gene expression of osteoclast-specific markers (TRAP, cathepsin K, MMP9) by quantitative PCR, and bone resorption activity by scanning electron microscopy. The purified ACPAs showed intact IgG structure of heavy and light chains, and specific binding to citrullinated vimentin, but not to native vimentin. However, ACPAs treatment did not significantly increase the number of TRAP-positive multinucleated cells, the expression of osteoclast-related genes, or bone resorption pit area compared to controls. Our findings demonstrate that pooled polyclonal ACPAs from RA patients, purified by MCEP affinity chromatography, do not directly promote osteoclastogenesis in vitro, suggesting that their osteoclastogenic effects may require a more complex microenvironment or depend on specific antibody properties not captured in this experimental setting.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s10238-026-02208-1.

Introduction

Rheumatoid arthritis (RA) is a chronic autoimmune disease characterized by persistent synovial inflammation and progressive joint destruction, a process in which osteoclast-mediated bone resorption plays a central role [1]. Anti-citrullinated peptide/protein antibodies (ACPAs) are highly specific serological markers for RA and contribute significantly to disease pathogenesis and progression [2,3]. Clinical studies have identified ACPA positivity as an independent predictor of radiographic progression, which is strongly correlated with the severity of bone erosion in RA [4–6].

Beyond their diagnostic utility, ACPAs are hypothesized to actively promote osteoclastogenesis. A foundational study by Harre et al. demonstrate that affinity purified IgG against mutated citrullinated vimentin (MCV) could bind specifically to osteoclast precursors and directly induce osteoclast differentiation and bone loss both in vitro and in vivo [7]. This provided a direct mechanistic link between the adaptive immunity and dysregulated bone metabolism in RA.

However, the direct osteoclastogenic effect of ACPAs has not been consistently replicated, revealing a more complex picture. A subsequent study proposing a chemokine-dependent mechanism for polyclonal ACPAs-induced osteoclastogenesis was later critically corrected due to a lack of specificity in the monoclonal antibodies used, highlighting the methodological challenges and the potential for unknown, non-ACPA-specific effects [8–10]. Furthermore, their direct contribution to bone erosion appears to be highly context dependent. The same group that first reported the pro-osteoclastogenic effect also found that in vitro sialylation of monoclonal ACPAs abrogated this ability [11]. Conversely, other studies have reported that certain monoclonal ACPAs can paradoxically attenuate arthritis severity and reduce bone erosion and cartilage damage in a mouse model of collagen antibodies induced arthritis (CAIA) [12–14].

These discrepancies suggest that the biological activity of ACPAs is likely modulated by factors such as purification methods, antibody subtype, IgG-Fc glycosylation patterns, and components of the in vivo microenvironment.

Consequently, the precise role of polyclonal ACPAs, as they naturally occur in patient sera, in directly driving osteoclast formation remains an open and debated question. This study therefore aims to determine whether polyclonal ACPAs, purified from RA patient sera, directly enhance osteoclast differentiation under standardized in vitro conditions.

Methods

Participants

This study enrolled patients with RA who were diagnosed and recruited from the Department of Rheumatology and Immunology at Peking University Third Hospital. RA patients were required to fulfill the 2010 American College of Rheumatology (ACR)/European League Against Rheumatism (EULAR) Classification criteria for RA [2]. Healthy controls were recruited from the same hospital during the same period. All participants provided written informed consent prior to enrollment. The study was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of Peking University Third Hospital (approval number: M20250032).

Purification of polyclonal ACPAs

Total IgG was initially isolated from pooled sera of 37 ACPA positive RA patients by Protein G affinity chromatography [15]. ACPA (+)-IgG was then purified from this IgG fraction using a multi-citrullinated-epitope peptide (MCEP) (Sequences: H-Ttds-AXAXGSGSGXGXG-NH2; X stands for citrulline; Ttds, linker 1,13-diamino-4,7,10-trioxatridecan-succinamic acid) conjugated affinity column, as previously described [15,16]. Control IgG sample were obtained from ACPA negative RA patients (n = 10; ACPA (-)-IgG) and a healthy donor (H-IgG). The detailed demographic characteristics of all participants are summarized in Table 1.

Table 1.

Clinical data on patients with RA and healthy control used for purification of IgG

Characteristics	ACPA + RA (n = 37)	ACPA- RA(n = 10)	P	HC(n = 1)
Age(years)	41.73(14.06)	45.8(13.07)	0.41	47
Sex	-	-	0.25	-
Male, n(%)	8(21.62)	4(40)	-	-
Female, n(%)	29(78.38)	6(60)	-	1(100)
Duration(months)	24(9.5, 90)	6(2.75, 24.75)	0.06*	-
CRP (mg/L)	3.2(1.7, 7.75)	2.2(1.83, 4.45)	0.63	-
ESR (mm/h)	13(6.5, 28)	6.5(2.75, 14.5)	0.07	-
DAS28-ESR	2.73(1.22)	3.49(1.56)	0.18	-
Anti-CCP (U/mL)	1145(289.5, 2500)	-	-	-
RF positive, (%)	30(81.08)	4(40)	0.02	-
Medication, n(%)	-	-	-	-
csDMARDs	28(75.68)	8(80)	> 0.99	-
bDMARDs/ tsDMARDs	11(29.73)	2(20)	0.70	-
Glucocorticoid	2(5.41)	0(0)	> 0.99	-
Bone erosion (%)	8/30(26.67)	1/9(11.11)	0.65	-

* Disease duration showed a borderline difference between groups (p = 0.06), with ACPA+ patients having longer disease duration. However, all disease activity measures (CRP, ESR, DAS28) were comparable between groups (all p > 0.05), with no consistent trend. Given the small number of ACPA- sample size (n = 10), this borderline p-value is unlikely to represent a clinically meaningful confounder for our primary findings

RA: rheumatoid arthritis

ACPA: anti-citrullinated peptide/protein antibodies

HC: healthy control

CRP: C-reactive protein

ESR: erythrocyte sedimentation rate

DAS28-ESR: disease activity score in 28 joints base on ESR

Anti-CCP: anti-cyclic citrullinated peptide antibody

RF: rheumatoid factor

csDMARDs: conventional synthetic disease-modifying antirheumatic drugs

b/tsDMARDs: biological/targeted synthetic DMARDs

SDS-PAGE and Coomassie Brilliant Blue staining

The recovery rate and purity of total IgG and ACPA (+)-IgG were analyzed by Sodium Dodecyl Sulphate-Polyacrylamide Gel Electrophoresis (SDS-PAGE) followed by Coomassie Brilliant Blue staining. Briefly, samples were diluted in 5× SDS loading buffer and denatured at 100 °C for 5 min. Serum, protein G column flow through (15 µg/lane), protein G column elution (both unpurified and purified), MCEP column flow through and elution (5 µg/lane) were loaded onto a 10% SDS-polyacrylamide gel. Electrophoresis was carried out at a constant voltage of 120 V for 90 min, and the gels were subsequently stained with Coomassie Brilliant Blue to visualize the proteins.

Western blot

The specificity of the purified ACPA (+)-IgG was evaluated by western blot. Vimentin and citrullinated vimentin (CAYMAN CHEMICAL, USA) (0.6 µg/lane) were separated by 10% SDS-PAGE and electrophoretically transferred onto polyvinylidene fluoride (PVDF) membranes (Merck Millipore). The membranes were blocked with 5% skimmed milk for 2 h at room temperature with gentle shaking, followed by overnight incubation at 4 °C with ACPA (+)-IgG, ACPA (-)-IgG, or H-IgG (8.24 µg/mL in 5 mL per group). Thereafter, the membranes were incubated with horseradish peroxidase (HRP)-conjugated goat anti-human IgG secondary antibody (1:10,000 dilution) for 1 h at room temperature. After thorough washing, the blots were visualized using an enhanced chemiluminescence (ECL) detection system.

Determination of anti-MCV antibodies by Enzyme-Linked Immunosorbent Assay (ELISA)

The concentrations of anti-MCV antibodies in purified IgG were measured with a commercial ELISA kit (Anti-MCV^®, ORGENTEC Diagnostika GmbH, Mainz Germany). All procedures were performed following the manufacturer’s instructions.

Determination of anti-CCP antibodies by Electrochemiluminescence Immunoassay (ECLIA)

The concentrations of Anti-CCP antibodies in purified IgG were measured with a commercial ECLIA assay (Cobas, Switzerland) following the manufacturer’s protocol in the clinical laboratory of Peking University Third Hospital.

Osteoclast culture

Peripheral blood mononuclear cells (PBMCs) were isolated within 2 h of collection from healthy donors and RA patients by density gradient centrifugation with Ficoll-Paque Plus (GE Healthcare, Sweden). CD14 + monocytes were then purified from PBMCs by magnetic activated cell sorting (Miltenyi Biotec, Germany) (purity > 99%). The fresh CD14 + monocytes were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated fetal bovine serum and 30 ng/mL M-CSF (PeproTech) for 3 days to induce osteoclast precursors differentiation. Osteoclastogenesis was induced by stimulating the precursors with 30 ng/mL M-CSF and 5 ng/mL RANKL (PeproTech), a concentration determined by preliminary dose-response experiments to be sub-optimal but sufficient for differentiation (Supplementary Figure S1), with or without the addition of IgG fractions (2 µg/mL). The culture medium was refreshed every 3 days. For a complete summary of RANKL concentrations used across all experiments (including pilot studies), see Supplementary Table S1.

TRAP staining

Upon 5–8 days of culture, when confluent large cells were observed under the microscope, the cells were fixed in 4% paraformaldehyde and stained for Tartrate resistant acid phosphatase (TRAP) using a leukocyte acid phosphatase kit (Sigma-Aldrich) according to the manufacturer’s protocol. TRAP-positive, multinucleated cells (≥ 3 nuclei) were counted as osteoclasts (OCs) under a light microscope. A minimum of 15 random microscopic fields per group were analyzed.

Immunofluorescence staining

After incubated with IgG fractions, cells were washed with PBS and fixed with 4% paraformaldehyde for 15 min. After permeabilization with 0.2% Triton X-100 for 10 min and blocking with 2% BSA in PBST (PBS containing 0.1% Tween-20) for 30 min, bound human IgG was detected using an Alexa Fluor^® 647-conjugated anti-human IgG antibody (1:1000 dilution) (Proteintech) incubated for 1 h. Subsequently, cells were stained with iFluor™ 488-conjugated phalloidin (Yeasen Biotechnology) for 30 min and DAPI for 5 min, respectively, with thorough PBS washes between each step. Finally, samples were imaged using a high-content screening (PerkinElmer, Operetta CLS).

Quantitative real-time PCR analysis

Total RNA was extracted using TRIzol reagent (Invitrogen). Subsequently, cDNA synthesis was performed using the FastKing RT Kit (TIANGEN BIOTECH). Target gene expression was analyzed by quantitative real-time PCR (qPCR) using Talent qPCR PreMix (SYBR Green) (TIANGEN BIOTECH) on a QuantStudio™ 5 Real-Time PCR System (Thermo Fisher Scientific). The primer sequences used were as follows: GAPDH, forward primer: 5’-ACCCAGAAGACTGTGGATGG-3’, reverse primer: 5’- CAGTGAGCTTCCCGTTCAG-3’; TRAP, forward primer: 5’-CCTA CCCACTGCCTGGTCAA-3’, reverse primer: 5’-ACGTAGCCCACGCCATTCTC-3’; Cathepsine k(CSK), forward primer: 5’-AAGCCAGACAACAGATTTCCATCAG-3’, reverse primer: 5’- ATAGCTCCCAGTGGGTGTCCAG-3’; MMP9, forward primer: 5’-GGGACGCAGACATCGTCATC-3’, reverse primer: 5’-TCGTCATCGTCGAA ATGGGC-3’.

Bone resorption assay

The CD14 + monocytes from both RA patients and healthy individuals were seeded and differentiated into osteoclasts on bovine bone slices (BioVendor) according to established methods. Following a 28-days culture period, the slices were fixed with 2.5% glutaraldehyde, dehydrated in an ethanol series, subjected to critical point drying, and sputter-coated with a gold layer prior to imaging with a scanning electron microscope (JEOL, JSM7900F, Japan). Resorption pits were imaged and their areas were quantified using ImageJ software, with a minimum of 10 random fields analyzed per group.

Statistical analysis

Statistical analyses were performed using GraphPad Prism software (version 9.5). Data with normal distribution (equal variance) are presented as mean ± standard deviation (SD) and were compared across groups by one-way ANOVA and Tukey’s multiple comparisons test. Non-normally distributed data are expressed as median [interquartile range, IQR] and compared using the Kruskal-Wallis test and Dunn’s multiple comparisons test. Given the small group sample sizes (n = 3/4), the normality tests had limited power. To ensure robust statistical inference, between-group comparisons were primarily performed using the Kruskal-Wallis test. A p value of less than 0.05 was considered statistically significant.

Results

Isolation, purification, and characterization of ACPAs from RA patients

Polyclonal ACPAs were purified from pooled sera of ACPA positive RA patients, using IgG from ACPA negative RA patients and a healthy individual as controls (Fig. 1a). As shown in Table 1, the ACPA positive and ACPA negative RA groups were well-matched in clinical features such as age, sex, disease duration, disease activity, and medication, with the exception of RF, which was more prevalent in the ACPA positive group. Analysis by SDS-PAGE and Coomassie Brilliant Blue staining of samples collected during purification revealed two clear bands corresponding to the heavy and light chains of IgG (Fig. 1b), with an estimated purity of > 96% as determined by densitometry. This indicated that effective removal of non-specific proteins by Protein G affinity chromatography and successful isolation of intact IgG molecules, with no notable protein degradation after dialysis and filtration. Western blot analysis further demonstrated that purified IgG from health control and RA patients did not bind native vimentin. In contrast, only ACPA (+)- IgG specifically recognized citrullinated vimentin, confirming that their selective binding specificity for the citrullinated form of vimentin (Fig. 1c). Consistent with this specificity, quantitative ELISA further revealed that the ACPA (+)-IgG preparation contained a high titer of anti-MCV antibody (94.8 U/mL). In contrast, control IgG contained low titers (H-IgG, 22.07U/mL; ACPA (-)-IgG, 14.08 U/mL) (Fig. 1d). Furthermore, the ACPA (+)-IgG also contained a high level of anti-CCP antibody (615 U/mL). As expected, anti-CCP antibodies were not detected in the control groups (clinical values < 8 U/mL). Together, these results establish the successful isolation of polyclonal ACPAs, providing a reliable basis for subsequent in vitro studies.

Fig. 1.

Isolation, purification, and characterization of ACPAs from RA patients. (a) Flowchart of Protein Purification; MCEP: multi-citrullinated-epitope peptide (b) Coomassie Brilliant Blue staining showing the recovery rate and purity of total IgG, ACPA (+)-IgG and ACPA (-)-IgG. Purification process ingredients ACPA (+)-IgG display in lane 1–6: ACPA positive serum (lane 1); Protein G column flow through (lane 2); Protein G column IgG eluate, unpurified (lane 3); Protein G column IgG eluate, purified (lane 4); MCEP column IgG flow through (lane 5); MCEP column IgG eluate (lane 6); Purification process ingredients ACPA (-)-IgG display in lane 7–10: ACPA negative serum (lane 7); Protein G column flow through (lane 8); Protein G column IgG eluate, unpurified (lane 9); Protein G column IgG eluate, purified (lane 10). (c) Western blot showing the binding of purified antibodies against vimentin (Vim) and citrullinated vimentin (Cit-Vim). (d) Detection of anti-MCV antibody concentration in the purified IgG by ELISA. The dashed line represents the clinical positive cut-off value of 20 U/mL. Figure 1a is created by BioGDP.com

Morphological analysis of osteoclasts by TRAP staining

To determine whether ACPA (+)-IgG directly promotes osteoclast differentiation, we evaluated osteoclasts formation by TRAP staining. The results clearly showed that the addition of purified ACPA (+)-IgG did not enhance osteoclast differentiation in CD14 + monocytes from either healthy donors or RA patients, irrespective of their ACPA status (Fig. 2a and b; Table 2). This lack of effect was corroborated by a dose-response experiment, which revealed no significant effect of ACPA (+)-IgG concentration on osteoclast differentiation (Fig. 2c). High-content imaging analysis revealed that the purified IgG from the three sources all bound to both mononuclear osteoclast precursors and multinucleate mature osteoclasts (Fig. 2d). No significant increase in the number of TRAP-positive multinucleated cells was detected, indicating that ACPAs do not directly promote osteoclastogenesis under the present in vitro conditions.

Fig. 2.

Morphological analysis of osteoclasts by TRAP staining. TRAP staining of OCs derived from CD14 + monocytes of (a) healthy individuals (n = 4) or (b) patients with RA (n = 4) on the day5-8 (Scale bar is 200 μm). (c) TRAP staining of OCs derived from CD14 + monocytes of healthy individual (n = 2), treated with varying concentrations of ACPA (+)-IgG, on the day8 (Scale bar is 200 μm). Immunofluorescence staining and analysis of purified IgG binding to osteoclast precursors and mature cells (d), red, staining for human IgG; green, phalloidin staining for actin; blue, DAPI staining for the nucleus. White arrowheads indicate staining for purified IgG (Scale bar is 50 μm). The bar graph shows the relative number of osteoclasts in each group compared to the M + R group. Data are presented as individual data points, with the bar height and error bars representing the median [IQR], respectively. Statistical analyses were conducted using Kruskal-Wallis test. M for M-CSF, R for RANKL, UG for µg/mL

Table 2.

Clinical data on people used for osteoclast culture

Number	Sex	Age	Duration (years)	CRP (mg/L)	ESR (mm/h)	DAS28-ESR	Anti-CCP (U/mL)	RF (IU/mL)	csDMARDs	bDMARDs/ tsDMARDs	Glucocorticoid	Bone erosion
RA1	F	58	18	5	57	6.44	116	< 20	MTX + LEF	No	No	No
RA2	F	68	18	13.5	58	6.19	179	< 20	IGU + HCQ+SSZ	No	No	No
RA3	F	67	10	0.62	13	2.79	< 8	< 20	No	No	No	Yes
RA4	F	75	6	1.81	45	6.00	124	1980	MTX + HCQ+IGU	ADA	Yes	Yes
RA5	F	71	8	0.25	12	2.04	< 8	< 20	LEF + HCQ+SSZ + IGU	No	Yes	-
RA6	F	70	36	0.65	55	3.88	65.5	30.2	TG	No	Yes	No
RA7	F	43	18	0.35	16	4.50	8885	213	No	No	No	Yes
RA8	F	59	3	22.4	30	4.68	197	63.7	No	No	No	No
N1	F	25	-	-	-	-	-	-	-	-	-	-
N2	M	40	-	-	-	-	-	-	-	-	-	-
N3	F	47	-	-	-	-	-	-	-	-	-	-
N4	F	38	-	-	-	-	-	-	-	-	-	-

RA: rheumatoid arthritis

CRP: C-reactive protein

ESR: erythrocyte sedimentation rate

DAS28-ESR: disease activity score in 28 joints base on ESR

Anti-CCP: anti-cyclic citrullinated peptide antibody

RF: rheumatoid factor

csDMARDs: conventional synthetic disease-modifying antirheumatic drugs

b/tsDMARDs: biological/targeted synthetic DMARDs

MTX: methotrexate

LEF: leflunomide

IGU: iguratimod

HCQ: hydroxychloroquine

SSZ: sulfasalazine

TG: thunder god vine, phytotherapeutic agents for rheumatism

ADA: adalimumab

Detection of osteoclast-specific markers expression during osteoclastogenesis

To further evaluate the effect of ACPA (+)-IgG on osteoclastogenesis, we analysis the mRNA expression of key osteoclast markers, including TRAP, Cathepsin K, and MMP9. Following treatment with ACPA (+)-IgG, quantitative PCR analysis revealed no significant upregulation in the expression of any these genes. As shown in Fig. 3a and b, the transcript levels were comparable to those in control groups, supporting the conclusion that ACPAs do not directly promote osteoclast differentiation at the transcriptional level.

Fig. 3.

Detection of osteoclast-specific markers expression during osteoclastogenesis. Transcription levels of TRAP, cathepsin K(CTSK), and MMP9 of OCs derived from CD14 + monocytes of (a) healthy individuals(n = 3) or (b) patients with RA(n = 3) on the day6. Data are presented as individual data points, with the bar height and error bars representing the median [IQR], respectively. Statistical analyses were conducted using Kruskal-Wallis test. M for M-CSF, R for RANKL

Functional analysis of osteoclast-mediated bone resorption

To assess the functional impact of ACPA (+)-IgG on osteoclast activity, we performed a bone resorption assay in which osteoclasts were cultured on bone slices. Resorption pits were visualized and quantified by scanning electron microscopy. The results indicated that no increase in resorption pit area in the presence of ACPA (+)-IgG relative to control groups (Fig. 4a-c). Given the limited number of biological replicates, these data are presented as descriptive observations without statistical inference. In this exploratory dataset, no obvious increase in resorption pit area was observed in the presence of ACPA (+)-IgG relative to control group. These preliminary findings do not support a pro-resorptive effect of ACPAs under the tested conditions, but they should be interpreted with caution and validated in larger cohorts.

Fig. 4.

Functional analysis of osteoclast-mediated bone resorption. Representative scanning electron microscopy images of bone resorption assay of (a) healthy individual(n = 1) and of (b) patients with RA (n = 2). (Scale bar is 10 μm). (c) The bar graph shows the ratio of bone resorption pit areas in each group compared to the M + R group based on analysis of 10 random fields per donor. Each data point represents one independent donor (healthy: n = 1; RA: n = 2). Data are presented as individual data points, with the bar height and error bars representing the median [IQR], respectively. No statistical tests were applied. M for M-CSF, R for RANKL

Discussion

In this study, polyclonal ACPAs were isolated and purified from pooled sera of RA patients using an established methodology. The purified ACPAs exhibited specific binding to citrullinated vimentin (including MCV) and were confirmed to be of high purity (> 96%). To evaluate their direct role in osteoclastogenesis, we employed a low RANKL concentration (5 ng/mL) in the osteoclast differentiation assay (Supplementary Figure S1), an approach adopted to avoid the robust signaling induced by high RANKL levels that could obscure the potential contribution of ACPAs. Consistent with our initial experiments using ACPAs from two patients with high anti-CCP titers (> 2500 U/mL) under high RANKL stimulation (100 ng/mL) (Supplementary Figure S2-S5), the pooled polyclonal ACPAs did not significantly promote osteoclast differentiation or bone resorption activity under these optimized conditions, as determined by morphological staining, osteoclast-specific genes expression analysis, and functional resorption assays. Critically, high-content imaging confirmed that polyclonal ACPAs significantly bound to both osteoclast precursors and mature cells (Fig. 2d), effectively ruling out a failure of target engagement as an explanation for the absence of a direct functional effect.

Therefore, despite their clear ability to recognize and bind to cells within the osteoclast lineage, pooled polyclonal ACPAs are insufficient to enhance the osteoclast differentiation in vitro.

In contrast (Supplementary Table S2), Harre et al. previously demonstrated that affinity-purified human autoantibodies against mutated citrullinated vimentin (MCV) strongly induced osteoclast differentiation and bone loss both in vitro and in vivo [7]. It is noteworthy, however, that their study also indicated that ACPAs did not enhance the bone resorption capacity of mature osteoclasts [7]. Although Krishnamurthy et al. later described a similar osteoclast-promoting effect [8], the authors themselves subsequently issued a correction stating that the specificity of their functional monoclonal antibodies could not be confirmed and emphasizing that conclusions based on polyclonal ACPAs data required cautious interpretation and further validation [9,10]. Similarly, previous work has described monoclonal ACPA with varying effects on osteoclastogenesis, including slight enhancement, inhibition, or no effect [17].

Our independent effort to replicate this key finding using purified ACPAs did not reveal a significant osteoclastogenic effect, thereby align with the more cautious perspective articulated in the correction by Krishnamurthy et al., and suggest that the capacity of ACPAs to directly promote osteoclastogenesis may not be a universal or readily reproducible phenomenon across all ACPA-positive IgG pools.

A critical consideration when interpreting our negative findings is the relationship between epitope specificity and osteoclastogenic potential. Previous studies have reported that ACPAs promoting osteoclastogenesis bound specific epitopes, such as cit-vimentin G46-60 (Harre et al.) and G2-17/60–75 (Krishnamurthy et al.) (Supplementary Table S2). However, whether this epitope is uniquely required for the osteoclastogenic effect, or whether ACPAs targeting other epitopes may also possess similar activity, remains unknown. In the present study, we used a distinct citrullinated multi-peptide sequence for affinity purification (Supplementary Table S2). The purified ACPAs using this novel matrix recognized multiple citrullinated peptides, including filaggrin19, type II collagen, vimentin peptide, EBNA-2, and α-enolase with varying intensities, demonstrating high cross-reactivity [16]. Our validation experiments confirmed its specific binding to full-length citrullinated vimentin (amino acids 2-466) and MCV. However, this approach does not permit identification of the precise epitope(s) recognized. Therefore, we cannot ascertain whether the specific osteoclast-reactive epitope(s) are present within our binding repertoire. This represents an inherent limitation, our ACPA pool may simply lack the specific clonal populations required for direct osteoclast activation, which could account for the absence of a functional effect in our assays.

A second consideration regarding our study design relates to the pooling strategy used for ACPA purification. Pooling sera from 37 patients, while necessary to obtain sufficient material, carries a theoretical risk of diluting potentially rare pathogenic ACPA clones. However, several lines of evidence argue against this concern. First, the use of pooled patient samples is a commonly used approach in the field. In the landmark study by Krishnamurthy et al. [8], functional polyclonal ACPAs were isolated from pooled synovial fluid (n = 25) and pooled peripheral blood (n = 35) of RA patients, and this strategy successfully yielded positive pro-osteoclastogenic effects. This demonstrates that serum pooling does not inherently prevent detection of functional activity. Second, we performed pilot experiments using ACPAs purified from smaller pools of two individual patient sera (Supplementary Figures S2–S5); these preparations also failed to show significant pro-osteoclastogenic effects, suggesting that the absence of effect is not unique to our large pool. Third, we used a high total IgG concentration (2 µg/ml) in all assays, which is substantially above the estimated levels of ACPAs in Harre’s and Krishnamurthy’s studies. Thus, while the theoretical possibility of dilution cannot be completely excluded, our empirical data from both large and small pools consistently show no effect. Future studies using ACPAs from individual patients with severe erosive disease could provide definitive confirmation, but the current evidence suggests that direct pro-osteoclastogenic ACPAs, if they exist, are not a general phenomenon.

Beyond the potential dilution effects of pooling discussed above, the discrepancy across studies points to more complex underlying causes. As summarized in the Supplementary Table S2, while technical variables such as pool size and antibody concentration may contribute to the discrepancy across studies, they are unlikely to fully account for the divergent outcomes. These divergent outcomes between our data and the positive reports by Harre et al. may be explained by several critical factors, especially the profound heterogeneity of ACPAs. Variables such as IgG-Fc glycosylation status, clonal composition, and the source of monocytes may influence their functional effects. For example, Harre et al. also found that pro-osteoclastogenic effect of ACPAs is attenuated by glycosylation [11,18]. As a theoretical limitation, sialylation status of the isolated ACPA was not assessed.

More broadly, within the polyclonal ACPAs pool we purified, different ACPA clones likely target distinct citrullinated epitopes and exhibit functional heterogeneity: some subtypes may possess direct pro-inflammatory or pro-osteoclastogenic potential, and such potentially pathogenic clones can be diluted below a functional threshold; others may act indirectly via immune complexes modulated by glycosylation [18,19], complement activation, or Fcγ receptor cross-linking [11,18]; while yet others may inhibit osteoclastogenesis, potentially resulting in an overall neutral net effect. Our present in vitro findings do not exclude the possibility that certain ACPA subsets may require immune complex formation or multivalent Fc receptor cross-linking to exert a robust pro-osteoclastogenic signal, as we used the soluble form. Alternatively, it is possible that all the ACPAs clones isolated in our polyclonal pool exerted a modest pro-osteoclastogenic effect that was difficult to reproduce under simplified in vitro culture conditions.

Therefore, this discrepancy provides important insight into the inherent complexity of ACPAs biological functions. Emerging evidence indicates that certain ACPA variants may even confer protective effects. For instance, the monoclonal antibody E4 specifically binds citrullinated antigen in articular cartilage, forms immune complexes with citrullinated α-enolase and citrullinated type II collagen, and subsequently interact with FCGR2B on M2 like macrophages, leading to increased IL-10 secretion and reduced osteoclasts formation [12].

Hence, our results support the view that the direct bone-eroding potential is a specialized property of particular ACPA subsets, rather than a generic feature of the overall ACPA response.

The absence of direct osteoclastogenic response to purified ACPAs in vitro does not preclude a significant pathogenic role mediated within the complex synovial joint microenvironment. In RA, multiple immune and stromal cells—such as fibroblast-like synoviocytes, macrophages and neutrophils—coexist [1]. ACPAs may stimulate these cells to release osteoclastogenic factors (e.g., M-CSF and RANKL) and pro-inflammatory cytokines (e.g., TNF-α, IL-1β, IL-6 and IL-8), thereby indirectly promoting osteoclastogenesis [1,20,21]. Moreover, ACPA-antigen immune complexes can amplify inflammation via Fcγ receptor and complement activation, ultimately contributing to bone destruction [22].

The role of ACPAs is also likely stage-dependent and context-specific. ACPAs appeared years before clinical onset, coupled with the dissociation between seropositivity and radiographic progression in some patients [23,24]. Protective epitopes might operate early in the disease but become overshadowed by pathogenic responses after the onset of symptoms, a concept that underscores incompletely elucidated phenotypes and temporal windows of ACPA action.

Our study has several other limitations. First, while our finding that pooled polyclonal ACPAs exert no net effect inherently points to underlying functional heterogeneity. A key interpretative limitation is that without parallel data from monoclonal ACPAs, we cannot deconvolute whether this net result reflects uniform weakness or a balance of opposing forces. Second, the experimental model used was relatively simplified and the function of pooled polyclonal ACPAs lacked validation in a fully integrated immune microenvironment. Specifically, the current bone resorption assay did not simulate key features of the RA synovial microenvironment, such as its specific cytokine profile or hypoxic conditions. Finally, the study did not delve into the underlying mechanisms. Whether ACPAs influences early signaling pathways such as NF-κB and ERK requires further investigation, even though they did not directly promote osteoclast differentiation. We also acknowledge that clinical heterogeneity among RA monocyte donors (e.g., treatment, disease activity) represents a theoretical limitation, though the small sample size precludes further stratification.

Conclusion

Our data indicate that pooled polyclonal ACPAs, purified by MCEP affinity chromatography, do not induce osteoclast formation in vitro. This work refines our understanding of ACPA pathogenicity, shifting the paradigm from a blanket pro-osteoclastogenic label toward a more nuanced understanding of functional heterogeneity. It also reframes the key question from whether ACPAs cause bone loss to which specific subsets are pathogenic and through what in vivo mechanisms they act. Future research should focus on deciphering the basis of ACPA functional heterogeneity using monoclonal or epitope-specific ACPAs, alongside elucidating their precise roles within the multicellular environment of the RA joint. Unravelling these mechanisms is essential for developing targeted therapeutic strategies to prevent bone erosion in ACPA positive RA.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1 ^{(1.3MB, pdf)}

Author contributions

Methodology: Chen, An, Li, Zhao; Formal analysis and investigation: Chen; Writing - original draft preparation: Chen; Writing - review and editing: Chen, Li, Zhao; Funding acquisition: Li, Zhao; Supervision: Li, Zhao.

Funding

This work was funded by the National Natural Science Foundation of China (No. 82271825), and Beijing Municipal Natural Science Foundation (No.7222264). The authors have no relevant financial or non-financial interests to disclose.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of Peking University Third Hospital (No. M20250032).

Consent to participate

Informed consent was obtained from all individual participants included in the study.