Diagnostic Value of Antibody Responses to Mycobacterium avium subsp. paratuberculosis -Derived Proteins PtpA and PtpB in Rheumatoid Arthritis

1 1 Diagnostic Value of Antibody Responses to 2 Mycobacterium avium subsp. paratuberculosis-Derived 3 Proteins PtpA and PtpB in Rheumatoid Arthritis 4 5 Jorge Hernández-Bello 1, Sergio Cerpa-Cruz 4, Gabriela A. Sánchez-Zuno 5, Ferdinando 6 Nicoletti 6, Horacio Bach 2,*, José F. Muñoz-Valle1,* 7 8 1 Instituto de Investigación en Ciencias Biomédicas, Centro Universitario de Ciencias de la 9 Salud Universidad de Guadalajara, 44340 Guadalajara, México 10 2 Division of Infectious Diseases, Faculty of Medicine, The University of British Columbia, 11 Vancouver, BC, V6H 3Z6, Canada 12 3 Division of Rheumatology, Guadalajara Civil Hospital "Fray Antonio Alcalde", 13 Guadalajara, Jalisco, México 14 4 Department of Medicine, Yale School of Medicine, New Haven, CT, USA 15 5 Department of Biomedical and Biotechnological Sciences, University of Catania, Catania, 16 Italy 17 18 Corresponding authors 19 * Horacio Bach, Division of Infectious Diseases, Faculty of Medicine, The University of 20 British Columbia, 410-2660 Oak Street, Vancouver, BC V6H 3Z6, Canada. E- 21 mail:[email protected]. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 2 22 * José F. Muñoz-Valle, Instituto de Investigación en Ciencias Biomédicas, Centro 23 Universitario de Ciencias de la Salud, Universidad de Guadalajara, 44340 Guadalajara, 24 México. E-mail: [email protected] 25 26 Abstract 27 Evidence suggests that Mycobacterium avium subspecies paratuberculosis (MAP) may 28 contribute to autoimmune diseases such as rheumatoid arthritis (RA), partly through effector 29 proteins—particularly the tyrosine phosphatases PtpA and PtpB—that modulate macrophage 30 signaling and promote bacterial persistence. This study evaluated whether serum antibodies 31 against these proteins serve as biomarkers of RA. Humoral responses to PtpA and PtpB were 32 quantified in Mexican RA patients (n = 100) and healthy controls (n = 100) using in-house 33 ELISAs. Associations with disease activity (DAS28), ROC performance, and logistic 34 regression models were assessed. Results showed that anti-PtpB antibody levels were 35 significantly higher in patients with RA than in healthy controls (median OD 0.185 vs. 0.080; 36 p < 0.0001) and had moderate discriminative capacity (AUC = 0.762). Anti-PtpB reactivity 37 increased with higher disease activity and showed a significant positive association with 38 DAS28 (p < 0.05). In addition, there was a functional disability measured by HAQ (p < 39 0.001), as well as moderate correlations with erythrocyte sedimentation rate and rheumatoid 40 factor. A combined logistic regression model integrating both antibodies markedly improved 41 diagnostic accuracy (AUC = 0.934), achieving high sensitivity (90%) and specificity (89%). 42 These findings support a potential role of MAP in RA immunopathogenesis and indicate that 43 combined quantification of anti-PtpA and anti-PtpB antibodies captures complementary and 44 non-redundant immunological information. This combined serological approach may .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 3 45 enhance RA diagnosis and provide clinically relevant insights into disease activity and 46 severity. 47 48 Keywords: Rheumatoid arthritis; Mycobacterium avium subspecies paratuberculosis; PtpA; 49 PtpB; DAS-28; Tyrosine phosphatases; ELISA; 50 51 Introduction 52 RA is a multifactorial autoimmune disease in which genetic predisposition, 53 dysregulated immune pathways, and microbial exposures interact to promote chronic 54 synovial inflammation and joint destruction [1,2]. Growing interest has focused on 55 microorganisms capable of persisting within host immune cells and generating antigenic 56 stimuli that may shape autoantibody production or amplify inflammatory cascades. Among 57 these, Mycobacterium avium subsp. paratuberculosis (MAP) has emerged as a plausible 58 environmental trigger of autoimmunity due to its ability to survive within macrophages and 59 modulate intracellular signaling via secreted virulence factors [3–5]. 60 MAP effector proteins involved in host–pathogen interactions have attracted 61 particular attention for their immunogenic properties and potential relevance in RA. A pivotal 62 study from Italy demonstrated that the MAP-derived protein tyrosine phosphatases PtpA and 63 PknG are recognized at significantly higher frequencies in the sera of RA patients than in 64 those of healthy controls, supporting the notion that MAP exposure may leave a detectable 65 humoral footprint in RA [6]. Building on this idea, our group recently reported that PtpA- 66 specific antibodies are also elevated in Mexican patients with RA and may serve as an 67 informative immunological marker in this population [7]. These observations collectively .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 4 68 suggest that MAP phosphatases and kinases are antigenic targets that can elicit differential 69 immune responses in RA. 70 Mechanistically, MAP-secreted proteins such as PtpA and PtpB may become relevant 71 to RA pathogenesis because their intracellular effects converge on processes central to joint 72 inflammation. PtpA inhibits phagolysosomal fusion by dephosphorylating the human 73 VPS33B, a protein involved in phagolysosome fusion. This dephosphorylation allows 74 bacteria to persist within macrophages [8], major producers of the pro-inflammatory 75 cytokines TNF, IL-1β, and IL-6, which drive synovitis and structural damage [9]. Persistent 76 MAP antigens could therefore act as chronic stimuli, maintaining macrophage activation, 77 promoting continuous cytokine release, and enhancing Th1/Th17 polarization [10]. In 78 genetically susceptible individuals, repeated exposure to these antigens may also increase 79 autoantibody formation via molecular mimicry [11]. Furthermore, a higher antigenic load or 80 stronger immune recognition of MAP phosphatases may reflect ongoing innate immune 81 activation, potentially explaining an association between elevated antibody levels and greater 82 clinical activity in RA. 83 Among MAP-secreted effectors, the tyrosine phosphatases PtpA and PtpB have 84 attracted attention for their ability to disrupt phagosomal maturation and phosphoinositide 85 metabolism, thereby interfering with vesicular trafficking and promoting intracellular 86 survival—mechanisms that have been extensively characterized in related mycobacterial 87 pathogens. Although evidence has begun to accumulate for PtpA-specific responses in RA, 88 whether PtpB elicits a similar or complementary humoral signature remains unknown. 89 Furthermore, whether combining immune responses to both phosphatases can improve 90 diagnostic discrimination has never been explored. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 5 91 In this study, we analyzed antibody reactivity to PtpA and PtpB in patients with RA 92 and healthy controls and assessed the diagnostic utility of each marker individually and in 93 combination. By integrating our previous findings [7] with new data on PtpB, we aimed to 94 clarify the immunological relevance of MAP-secreted phosphatases in RA and determine 95 whether their combined measurement enhances diagnostic accuracy. 96 97 Materials and methods 98 99 Subjects 100 Archived serum samples from patients and healthy controls were used in this study. 101 These samples were used to determine the level of anti-PtpA in our previous study [7]. 102 Briefly, RA patients (23 males, 77 females; median age 58) who fulfilled the 2010 103 ACR/EULAR Classification Criteria for RA were enrolled at the Rheumatology Unit of the 104 Civil Hospital of Guadalajara, Fray Antonio Alcalde, Guadalajara, Jalisco, Mexico, between 105 January 1, 2018, and December 31, 2021. Clinical and demographic data were collected, 106 including disease duration, rheumatoid factor (RF) and anti-cyclic citrullinated peptide (anti- 107 CCP) status, treatment with steroids and disease-modifying anti-rheumatic drugs 108 (DMARDs), C-reactive protein (CRP) levels, erythrocyte sedimentation rate (ESR), Disease 109 Activity Score-28 (DAS-28), and Health Assessment Questionnaire (HAQ) scores. 110 A group of 100 healthy controls (20 males, 80 females; median age 40 years) was 111 recruited at the same hospital. Control participants verbally confirmed having no prior history 112 of tuberculosis. Antibody reactivity to PtpA in this cohort has been previously described [7]. .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 6 113 For the present study, the same archived sera were additionally evaluated for reactivity to 114 PtpB, and combined analyses of PtpA and PtpB were performed. 115 116 Ethics approval statement 117 This study was approved by the Ethics Committee of the University of Guadalajara 118 (Approval No. 0122017) and conducted in accordance with the ethical principles outlined in 119 the Declaration of Helsinki (64th World Medical Association General Assembly, Fortaleza, 120 Brazil, 2013). Written informed consent was obtained from all participants prior to inclusion 121 in the study. 122 123 ELISA assays 124 Plate preparation 125 The recombinant PtpB protein was expressed in Escherichia coli harboring the ptpB 126 gene in the ampicillin-resistant pET-22 vector. Purification was performed via Ni-NTA 127 affinity chromatography, and the protein was stored at –20°C until use. The preparation of 128 recombinant PtpA and the corresponding assay conditions have been previously described 129 [7]. 130 For ELISA, Maxisorp plates (ThermoFisher) were coated with 50 μg/mL of antigen 131 in phosphate-buffered saline (PBS) and incubated overnight at 4°C. Plates were washed three 132 times with PBS containing 0.05% Tween-20 (PBS-T) and blocked with 3% bovine serum 133 albumin (BSA) in PBS at 4°C overnight. Plates were air-dried prior to use. 134 135 Assay procedure .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 7 136 Sera from RA patients and controls were tested in triplicate. After incubation with 137 sera, plates were washed with PBS-T and subsequently incubated with a peroxidase- 138 conjugated anti-human IgG secondary antibody. Optical density (OD) was measured at 450 139 nm using an Epoch microplate reader (BioTek, USA). The baseline signal, defined as the 140 secondary antibody alone, was subtracted from all readings. Positive controls were included 141 in all assays. Cut-off values were determined by Receiver Operating Characteristic (ROC) 142 analysis to ensure specificity above 90%, with sensitivity adjusted accordingly. 143 144 Statistical analysis 145 Differences in antibody reactivity between RA patients and controls were assessed 146 using the Mann–Whitney U test. Associations between clinical variables and antibody levels 147 were examined by linear regression. ROC curves and areas under the curve (AUC) were 148 generated using Python (version 3.14) with the scikit-learn library. A multivariable logistic 149 regression model integrating anti-PtpA and anti-PtpB antibody levels was constructed to 150 assess combined diagnostic performance. Pairwise comparisons between ROC curves were 151 performed using DeLong’s test. Sensitivity, specificity, positive predictive value (PPV), and 152 negative predictive value (NPV) were calculated at the optimal cutoff determined by the 153 Youden index. Graphical representations of ROC curves and correlation plots were generated 154 using GraphPad Prism version 8.0 (GraphPad Software, San Diego, CA). Statistical 155 significance was defined as a two-sided p-value < 0.05. 156 157 Results .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 8 158 Patients with RA were significantly older than controls (median 54 vs. 40 years, p 160 0.90). Among RA patients, disease activity and disability indices showed median DAS-28 161 and HAQ scores of 3.2 and 0.75, respectively. Inflammatory markers were elevated, with C- 162 reactive protein (CRP) at 6 mg/dL and ESR at 20 mm/h. Most patients were receiving 163 DMARDs (73%), and more than one-third reported NSAID use (36%), whereas only 1% 164 were on corticosteroid therapy (Table 1). 165 As shown in Fig 1A, antibody levels against PtpB were higher in the RA group than 166 in controls (p < 0.0001, Mann–Whitney U test). The median optical density (OD) in RA 167 patients was 0.1847 [25 th–75th percentile: 0.1120–0.2483], whereas the control group 168 exhibited a median OD of 0.0801 [25 th–75th percentile: 0.03275–0.1434]. ROC analysis 169 demonstrated that anti-PtpB antibodies effectively discriminated RA patients from healthy 170 controls, with an AUC of 0.762 (p < 0.0001; Fig 1B). 171 172 Fig 1. Serum antibody reactivity to PtpB in RA subjects and CS. (A) Comparison of anti- 173 PtpB antibody levels between groups. Bars represent the median and interquartile range; 174 dashed lines indicate antibody positivity thresholds, and p-values are shown above. (B) 175 Receiver operating characteristic (ROC) curve evaluating the discriminative capacity of anti- 176 PtpB antibodies. 177 178 To assess associations between anti-PtpB levels and disease activity, RA patients 179 were stratified into four groups according to their DAS28 scores: remission, low disease 180 activity, moderate disease activity, and high disease activity. As shown in Fig 2, patients in 181 remission had the lowest antibody levels (median = 0.1187; IQR = 0.0447–0.1887). Those .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 9 182 with low disease activity showed slightly higher values (median = 0.1752; IQR = 0.1053– 183 0.2296), with no significant difference relative to remission (p = 0.3209). The moderate 184 activity group displayed intermediate values (median = 0.1948; IQR = 0.1574–0.2527), 185 which overlapped with those of the low activity group (p = 0.9999). Conversely, patients 186 with high disease activity exhibited the highest antibody levels (median = 0.2357; IQR = 187 0.1682–0.3324) and differed significantly from the remission group (p = 0.0001). 188 189 Fig 2. Association between serum anti-PtpB antibody levels and DAS-28 categories. 190 Bars represent median ± interquartile range for each disease activity group. Statistical 191 analysis was performed using the Kruskal–Wallis test followed by Dunn’s post hoc 192 correction. P-values are shown above the distributions. 193 194 Fig 3 displays the correlation heatmap between anti-PtpB antibody levels and clinical 195 and laboratory variables in patients with RA. Anti-PtpB antibody levels showed statistically 196 significant positive correlations with disease activity and functional impairment, including 197 DAS28 (ρ = 0.45, p < 0.001) and HAQ (p = 0.40, p < 0.001). In addition, moderate positive 198 associations were observed with ESR (p = 0.37, p = 0.003) and rheumatoid factor (RF; p = 199 0.49, p < 0.001). No significant differences in anti-PtpB levels were observed by sex or 200 treatment status, including use of non-steroidal anti-inflammatory drugs (NSAIDs), 201 corticosteroids, sulfasalazine, chloroquine, or methotrexate. No other clinical, hematological, 202 or demographic variables were significantly associated with anti-PtpB levels. 203 204 Fig 3. Correlation heatmap of anti-PtpB antibody levels with clinical and laboratory 205 parameters in patients with RA. The heatmap displays Spearman correlation coefficients .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 10 206 (ρ) between anti-PtpB antibody levels and clinical and laboratory variables, including 207 hematological parameters, inflammatory markers, functional disability, and disease activity 208 indices. Each cell shows the corresponding correlation coefficient and p-value. Warm colors 209 indicate positive correlations, whereas cool colors indicate negative correlations. 210 Abbreviations: WBC, white blood cells; RBC, red blood cells; Hb, hemoglobin; MCV, mean 211 corpuscular volume; PLT, platelets; ESR, erythrocyte sedimentation rate; CRP, C-reactive 212 protein; RF, rheumatoid factor; BMI, body mass index; HAQ, Health Assessment 213 Questionnaire; DAS28, Disease Activity Score 28. 214 215 The relationship between humoral immune responses to the MAP-derived proteins 216 PtpA and PtpB was assessed by comparing antibody levels in RA patients and controls. As 217 shown in Fig 4A and 4B, no significant correlation was found between anti-PtpA and anti- 218 PtpB levels in either group. 219 220 Fig 4. Correlation between anti-PtPA and anti-PtpB antibody levels in RA patients and 221 control subjects. (A) RA patients. (B) Control subjects. Spearman’s rank correlation 222 analysis was used to assess associations between markers. Blue and red lines denote the best- 223 fit linear regressions for each group. 224 225 To evaluate the diagnostic utility of combining anti-PtpA and anti-PtpB antibodies, a 226 multivariable logistic regression model was constructed and its performance assessed. As 227 shown in Table 2, the combined model demonstrated excellent discriminative capacity, with 228 an AUC of 0.934. At the optimal cut-off value of 0.3869, determined using the Youden index, 229 the model achieved a sensitivity of 96% and a specificity of 87%. This threshold also yielded .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 11 230 high predictive values, with a positive predictive value (PPV) of 86% and a negative 231 predictive value (NPV) of 97%. 232 ROC curves for anti-PtpA, anti-PtpB, and their combination are presented in Fig 5. 233 The combined model outperformed both individual markers (PtpA AUC = 0.925; PtpB AUC 234 = 0.762). Pairwise comparisons of ROC curves using DeLong’s test revealed no statistically 235 significant differences between the combined model and anti-PtpA alone (ΔAUC = 0.009, p 236 = 0.975), nor between the combined model and anti-PtpB alone (ΔAUC = 0.172, p = 0.495). 237 Similarly, the difference between anti-PtpA and anti-PtpB was not statistically significant 238 (ΔAUC = 0.163, p = 0.485). 239 240 Fig 5. ROC curve comparison of models using anti-PtpA, anti-PtpB, and their 241 combination. The green solid line represents the combined PtpA+PtpB model, which 242 achieved the highest diagnostic accuracy. The blue dashed line represents the model using 243 only PtpA, while the orange dashed line represents the model using only PtpB. The black 244 diagonal line indicates random classification (AUC = 0.5). 245 246 Discussion 247 Emerging evidence suggests that bacterial exposures may contribute to the etiology 248 of RA by disrupting immune tolerance and promoting chronic inflammation in genetically 249 susceptible individuals [18]. Several microorganisms, particularly those capable of persisting 250 within macrophages or modifying host proteins, have been implicated as potential triggers of 251 autoimmunity [19]. Intracellular bacteria such as MAP can interfere with phagosome 252 maturation and sustain pro-inflammatory cytokine production [20,21]. Meanwhile, mucosal .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 12 253 pathogens, including Porphyromonas gingivalis and Aggregatibacter 254 actinomycetemcomitans, have been associated with aberrant protein citrullination and the 255 induction of anti-CCP antibodies in RA [22], supporting the concept that microbial factors 256 may converge on shared immunopathogenic pathways. 257 In this study, we evaluated humoral immune responses to the MAP-derived tyrosine 258 phosphatases PtpA and PtpB in patients with RA and healthy controls. Four principal 259 findings emerged: (i) anti-PtpB antibodies were significantly elevated in RA; (ii) anti-PtpB 260 titers increased with higher disease activity; (iii) anti-PtpA and anti-PtpB responses were 261 immunologically independent; and (iv) a combined logistic model incorporating both 262 antibodies markedly improved diagnostic accuracy. Collectively, these observations 263 strengthen the hypothesis that MAP antigens may contribute to the immunological landscape 264 of RA and may serve as complementary biomarkers in this population. 265 Our findings extend existing evidence suggesting MAP exposure in RA. Previous 266 studies have shown that MAP-secreted proteins, such as PtpA and PknG, are more frequently 267 recognized by RA sera than by healthy control sera [12]. We also previously reported 268 increased anti-PtpA responses in Mexican RA patients [7]. Molecular investigations in the 269 USA, Europe, and the Middle East have detected MAP DNA or MAP-reactive antibodies in 270 RA populations, reinforcing the plausibility of MAP as an environmental contributor to 271 autoimmunity [6,23–25]. 272 The present study adds the observation that PtpB, an established virulence 273 phosphatase in mycobacteria, also elicits increased antibody responses in RA. Given that 274 PtpB participates in immune evasion and intracellular persistence [16,26], heightened 275 seroreactivity in RA patients is biologically compatible with chronic antigen exposure. 276 Importantly, anti-PtpB antibody levels were positively associated with established measures .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 13 277 of disease activity and functional impairment, including DAS28 and HAQ, providing a 278 clinically meaningful link between MAP-related immune responses and disease expression. 279 The association between anti-PtpB antibody levels and clinical disease activity 280 provides a biologically plausible link between MAP-related immune responses and 281 established pathogenic mechanisms in RA. Macrophages play a central role in RA synovitis, 282 acting as key effector cells that produce pro-inflammatory cytokines such as TNF-α, IL-1β, 283 and IL-6, which drive both joint inflammation and structural damage [27]. MAP 284 phosphatases directly modulate macrophage biology: PtpA inhibits recruitment of the 285 vacuolar H⁺-ATPase to the phagosome, blocking acidification [28], whereas PtpB in other 286 mycobacteria modulates intracellular kinase signaling and promotes bacterial survival 287 [29,30]. In this context, elevated anti-PtpB titers in RA may reflect sustained activation of 288 innate immune pathways, consistent with their correlations with DAS28 and HAQ. 289 Anti-PtpB levels were also moderately associated with ESR and RF, but not with CRP 290 or BMI. This pattern suggests that anti-PtpB immunity does not simply mirror acute-phase 291 inflammation or metabolic status. Rather, it supports the existence of a more specific 292 immunological axis, potentially centered on macrophage activation and humoral 293 autoimmunity, in which MAP-related antigens contribute to disease severity without acting 294 as nonspecific inflammatory markers. The lack of association with CRP may further indicate 295 that anti-PtpB antibodies capture chronic or cumulative immune activation rather than 296 transient inflammatory fluctuations. 297 The absence of significant differences in anti-PtpB levels according to sex or 298 exposure to conventional antirheumatic therapies adds an important dimension to this 299 interpretation. These findings suggest that anti-PtpB responses are relatively stable and not .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 14 300 readily modulated by therapies that primarily target downstream inflammatory cascades, 301 raising the possibility that anti-PtpB antibodies reflect upstream or parallel pathogenic 302 processes that are not fully addressed by current therapeutic strategies. 303 One of the most distinctive findings was the independence between anti-PtpA and 304 anti-PtpB titers in both RA patients and controls. This observation is consistent with and 305 provides mechanistic support for our previous findings that anti-PtpA antibody levels were 306 not associated with DAS28, autoantibody status, or inflammatory markers in RA [7]. Such 307 independence likely reflects the functional divergence between the two phosphatases. PtpA 308 disrupts the VPS33B–V-ATPase axis, primarily affecting phagosomal maturation and 309 intracellular trafficking [31]. In contrast, PtpB operates through distinct lipid-mediated and 310 kinase-dependent signaling pathways that influence host immune activation [32,33]. Their 311 non-overlapping virulence mechanisms provide a plausible biological basis for differential 312 immune recognition, suggesting engagement of distinct antigen-processing routes and B-cell 313 activation pathways rather than a shared humoral response. In this framework, anti-PtpA 314 responses may reflect exposure-related or host–pathogen interactions, whereas anti-PtpB 315 immunity appears more closely linked to clinically relevant inflammatory and disease- 316 activity pathways. 317 The strong diagnostic performance of the combined anti-PtpA/anti-PtpB logistic 318 regression model further supports this complementary behavior. The model achieved 319 excellent discriminative accuracy (AUC = 0.934), a level conventionally interpreted as 320 indicative of high diagnostic utility [34]. Although pairwise ROC comparisons did not 321 demonstrate statistically significant superiority over individual antibodies, the combined 322 model consistently yielded numerically higher performance metrics. At the Youden- 323 optimized cut-off, the model achieved very high sensitivity (96%) and negative predictive .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 15 324 value (97%), while maintaining favorable specificity (87%) and positive predictive value 325 (86%). This diagnostic profile is comparable to that reported for multiepitope serological 326 tools used in early RA [35,36], underscoring the potential of MAP-derived antigens as 327 clinically meaningful complementary biomarkers rather than standalone diagnostic 328 replacements. 329 Accordingly, PtpA/PtpB serology is not intended to replace established RA 330 biomarkers such as RF or anti-citrullinated protein antibodies, which remain central to the 331 ACR/EULAR classification criteria. Instead, these MAP-derived immune markers may 332 provide orthogonal information related to disease biology and immune activation. 333 Significantly, future studies should extend the evaluation of these markers to other 334 autoimmune and inflammatory diseases to determine their disease specificity. 335 MAP has been implicated in autoimmune diseases through mechanisms of molecular 336 mimicry, including shared epitopes between MAP Hsp65 and human GAD65 [37]. Although 337 this study did not evaluate epitope overlap, cross-reactive immunity may contribute to the 338 amplification of adaptive responses, including the formation of RA-associated autoantibodies 339 [38]. 340 While this study was not designed to investigate transmission routes, previous work 341 has demonstrated that MAP can be acquired through consumption of unpasteurized dairy 342 products or contact with infected livestock, both of which have been associated with 343 increased MAP positivity in humans. Given that MAP is shed in milk, feces, and aerosols 344 from infected ruminants, these findings underscore the importance of zoonotic and foodborne 345 exposure pathways in interpreting MAP-derived immune responses [39,40]. Such reservoirs 346 may be particularly relevant in regions where traditional dairy practices persist or where rural 347 populations have greater exposure to livestock. Future studies should therefore incorporate .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 16 348 detailed exposure histories. In addition, the influence of long-term immunosuppressive 349 therapies should be carefully considered, as disease-modifying antirheumatic drugs may alter 350 host immune surveillance. 351 Some limitations should be considered when interpreting these findings. The cross- 352 sectional design does not allow causal relationships to be established, and direct detection of 353 MAP in tissues was beyond the scope of the present study. Although age differences were 354 observed between groups, age did not correlate with antibody titers, suggesting that 355 cumulative environmental exposure is unlikely to account for the observed MAP-reactive 356 immune responses. Together, these data provide a strong rationale for future longitudinal and 357 mechanistic studies to define further the pathogenic relevance of MAP-derived immune 358 signatures in RA and their potential utility for patient stratification. 359 Conclusions 360 This study demonstrates that anti-PtpB antibodies are elevated in RA, are associated 361 with disease activity and functional impairment, and provide clinically relevant information 362 that complements anti-PtpA responses. While anti-PtpA remains the strongest individual 363 discriminator between RA patients and healthy controls, anti-PtpB antibodies appear to 364 capture a distinct immunological dimension linked to inflammatory burden and disease 365 severity. Accordingly, the combined assessment of both antibodies achieves excellent overall 366 diagnostic performance, reflecting their complementary and non-redundant biological roles. 367 Importantly, anti-PtpB antibody levels were independent of sex and conventional 368 antirheumatic treatments, supporting the notion that MAP-related immune responses are not 369 merely secondary to therapy or demographic factors. Together, these findings reinforce the 370 hypothesis that exposure to MAP may represent a relevant environmental component in RA .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 17 371 immunopathogenesis, with different MAP-derived antigens contributing heterogeneously to 372 disease expression. 373 Further studies incorporating longitudinal sampling, tissue-level MAP detection, 374 treatment stratification, and host genetic factors are warranted to clarify the mechanistic and 375 clinical significance of MAP-related immunity in RA. Such efforts will also be essential for 376 determining the specificity and broader relevance of combined anti-PtpA/anti-PtpB profiling 377 in other autoimmune and rheumatic diseases in which MAP has been proposed as a potential 378 environmental trigger. 379 380 Acknowledgments 381 The authors would like to thank the patients who participated in this study. 382 383 Funding source 384 The Universidad de Guadalajara supported the work performed in México through the 385 Programa de Fortalecimiento de Institutos, Centros y Laboratorios de Investigación 2025. 386 The work performed in Canada was supported by the Antibody Engineering and Proteomics 387 Facility, Immunity and Infection Research Centre, Vancouver, Canada. The Universidad 388 Politécnica del Centro, Tabasco, México supported LAB. 389 390 Author contributions 391 JHB: Formal analysis, Investigation, Methodology, Validation, Visualization, Writing – 392 original draft, Writing – review and editing. HB: Conceptualization, Data curation, Formal 393 analysis, Investigation, Methodology, Validation, Visualization, Writing – original draft, .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 18 394 Writing – review and editing. SCC: Formal analysis, Investigation, Validation, Writing – 395 original draft, Writing – review and editing. GSZ: Formal analysis, Investigation, 396 Methodology, Validation, Visualization, Writing – review and editing. FN: Formal analysis, 397 Investigation, Methodology, Validation, Visualization, Writing – review & editing. FMV: 398 Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, 399 Visualization, Writing – original draft, Writing – review and editing. 400 401 Declaration of competing interest 402 The authors declare that they have no competing interests. 403 404 Data availability 405 Data will be made available upon reasonable request. 406 407 Supporting information 408 S1 Data. (XLSX) 409 410 References 411 1. Li J, Kuhn KA. Microbial threads in the tapestry of rheumatoid arthritis. J Clin Invest. 412 2025;135: e195374. doi:10.1172/JCI195374 413 2. Masoumi M, Solaymani M, Abbasifard M, Houshmandfar S, Iravani P, Saeedi- 414 Boroujeni A, et al. The genetic puzzle of rheumatoid arthritis: Causes, progression, 415 and treatment. Biochem Biophys Rep. 2025;43: 102148. 416 doi:10.1016/j.bbrep.2025.102148 417 3. Asgari N, Ghaemi EA, Tavasoli S, Aghaei M, Razavi Nikoo H, Sechi LA, et al. 418 Detection of Mycobacterium avium subspecies paratuberculosis in the blood of 419 patients with rheumatoid arthritis by using serological and molecular techniques. Curr 420 Microbiol. 2023;80: 359. doi:10.1007/s00284-023-03477-9 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 19 421 4. Asgari N, Ghaemi EA, Tavasoli S, Aghaei M, Nikoo HR, Zamani S. Exploring the 422 association between Mycobacterium avium subspecies paratuberculosis infection and 423 rheumatoid arthritis: an immunological perspective. Arthritis Res Ther. 2025;27: 36. 424 doi:10.1186/s13075-025-03501-2 425 5. Johansen MD, de Silva K, Plain KM, Whittington RJ, Purdie AC. Mycobacterium 426 avium subspecies paratuberculosis is able to manipulate host lipid metabolism and 427 accumulate cholesterol within macrophages. Microb Pathogen. 2019;130: 44–53. 428 doi:10.1016/j.micpath.2019.02.031 429 6. Bo M, Erre GL, Bach H, Slavin YN, Manchia PA, Passiu G, et al. PtpA and PknG 430 proteins secreted by Mycobacterium avium subsp. paratuberculosis are recognized by 431 sera from patients with rheumatoid arthritis: A case-control study. J Inflamm Res. 432 2019;12: 301–308. doi:10.2147/JIR.S220960 433 7. Hernández-Bello J, Bach H, Cerpa-Cruz S, Sánchez-Zuno GA, Hernández-Gutiérrez 434 R, Nicoletti F, et al. PtpA protein from Mycobacterium avium subsp. paratuberculosis 435 as a potential marker of rheumatoid arthritis in humans. PLoS One. 2025;20: 436 e0316727. doi:10.1371/journal.pone.0316727 437 8. Frehel C, de Chastellier C, Lang T, Rastogi N. Evidence for inhibition of fusion of 438 lysosomal and prelysosomal compartments with phagosomes in macrophages infected 439 with pathogenic Mycobacterium avium. Infect Immun. 1986;52: 252–262. 440 doi:10.1128/iai.52.1.252-262.1986 441 9. Fujimoto S, Niiro H. Pathogenic role of cytokines in rheumatoid arthritis. J Clin Med. 442 2025;14: 6409. doi:10.3390/jcm14186409 443 10. Ríos-Osorio N, Caviedes-Bucheli J, Sosa-Gutierrez J, Fernández-Grisales R. From 444 quiescence to proliferation: Molecular mechanisms driving Malassez epithelial cell 445 activation and proliferation into inflammatory radicular cystic transformation. Int 446 Endod J. 2025. doi:10.1111/iej.70062 447 11. Cusick MF, Libbey JE, Fujinami RS. Molecular mimicry as a mechanism of 448 autoimmune disease. Clin Rev Allergy Immunol. 2012;42: 102–111. 449 doi:10.1007/s12016-011-8294-7 450 12. Bach H, Sun J, Hmama Z, Av-Gay Y. Mycobacterium avium subsp. paratuberculosis 451 PtpA is an endogenous tyrosine phosphatase secreted during infection. Infect Immun. 452 2006;74: 6540–6546. doi:10.1128/IAI.01106-06 453 13. Bach E, Raizman EA, Vanderwal R, Soto P, Chaffer M, Keefe G, et al. 454 Immunogenicity of PtpA secreted during Mycobacterium avium ssp. paratuberculosis 455 infection in cattle. Vet Immunol Immunopathol. 2018;198: 1–5. 456 doi:10.1016/j.vetimm.2018.02.006 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 20 457 14. Rees WD, Lorenzo-Leal AC, Steiner TS, Bach H. Mycobacterium avium subspecies 458 paratuberculosis infects and replicates within human monocyte-derived dendritic 459 cells. Microorganisms. 2020;8: 994. doi:10.3390/microorganisms8070994 460 15. Grundner C, Ng H-L, Alber T. Mycobacterium tuberculosis protein tyrosine 461 phosphatase PtpB structure reveals a diverged fold and a buried active site. Structure. 462 2005;13: 1625–1634. doi:10.1016/j.str.2005.07.017 463 16. Mascarello A, Mori M, Chiaradia-Delatorre LD, Menegatti ACO, Monache FD, 464 Ferrari F, et al. Discovery of Mycobacterium tuberculosis protein tyrosine 465 phosphatase B (PtpB) inhibitors from natural products. PLoS One. 2013;8: e77081. 466 doi:10.1371/journal.pone.0077081 467 17. Li J, Kuhn KA. Microbial threads in the tapestry of rheumatoid arthritis. J Clin Invest. 468 135: e195374. doi:10.1172/JCI195374 469 18. Qiu CC, Caricchio R, Gallucci S. Triggers of autoimmunity: The role of bacterial 470 infections in the extracellular exposure of lupus nuclear autoantigens. Front Immunol. 471 2019;10: 2608. doi:10.3389/fimmu.2019.02608 472 19. Wherry TLT, Heggen M, Shircliff AL, Mooyottu S, Stabel JR. Stage of infection with 473 Mycobacterium avium subsp. paratuberculosis impacts expression of Rab5, Rab7, and 474 CYP27B1 in macrophages within the ileum of naturally infected cows. Front Vet Sci. 475 2023;10: 1117591. doi:10.3389/fvets.2023.1117591 476 20. Park H-T, Park H-E, Shim S, Kim S, Shin M-K, Yoo HS. Epithelial processed 477 Mycobacterium avium subsp. paratuberculosis induced prolonged Th17 response and 478 suppression of phagocytic maturation in bovine peripheral blood mononuclear cells. 479 Sci Rep. 2020;10: 21048. doi:10.1038/s41598-020-78113-8 480 21. Gómez-Bañuelos E, Mukherjee A, Darrah E, Andrade F, Gómez-Bañuelos E, 481 Mukherjee A, et al. Rheumatoid arthritis-associated mechanisms of Porphyromonas 482 gingivalis and Aggregatibacter actinomycetemcomitans. J Clin Med. 2019;8: 1309. 483 doi:10.3390/jcm8091309 484 22. Asgari N, Ghaemi EA, Tavasoli S, Aghaei M, Nikoo HR, Zamani S. Exploring the 485 association between Mycobacterium avium subspecies paratuberculosis infection and 486 rheumatoid arthritis: an immunological perspective. Arthritis Res Ther. 2025;27: 36. 487 doi:10.1186/s13075-025-03501-2 488 23. Asgari N, Ghaemi EA, Tavasoli S, Aghaei M, Razavi Nikoo H, Sechi LA, et al. 489 Detection of Mycobacterium avium Subspecies paratuberculosis in the blood of 490 patients with rheumatoid arthritis by using serological and molecular techniques. Curr 491 Microbiol. 2023;80: 359. doi:10.1007/s00284-023-03477-9 492 24. Sharp RC, Beg SA, Naser SA. Polymorphisms in protein tyrosine phosphatase non- 493 receptor Type 2 and 22 (PTPN2/22) are linked to hyper-proliferative T-cells and .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 21 494 susceptibility to Mycobacteria in rheumatoid arthritis. Front Cell Infect Microbiol. 495 2018;8: 11 doi:10.3389/fcimb.2018.00011 496 25. Mascarello A, Mori M, Chiaradia-Delatorre LD, Menegatti ACO, Delle Monache F, 497 Ferrari F, et al. Discovery of Mycobacterium tuberculosis protein tyrosine 498 phosphatase B (PtpB) inhibitors from natural products. PLoS One. 2013;8: e77081. 499 doi:10.1371/journal.pone.0077081 500 26. Fujimoto S, Niiro H, Fujimoto S, Niiro H. Pathogenic role of cytokines in rheumatoid 501 arthritis. J Clin Med. 2025;14: 6409. doi:10.3390/jcm14186409 502 27. Bach H, Papavinasasundaram KG, Wong D, Hmama Z, Av-Gay Y. Mycobacterium 503 tuberculosis virulence is mediated by PtpA dephosphorylation of human vacuolar 504 protein sorting 33B. Cell Host Microbe. 2008;3: 316-322. doi: 505 10.1016/j.chom.2008.03.008. 506 28. Beresford NJ, Mulhearn D, Szczepankiewicz B, Liu G, Johnson ME, Fordham- 507 Skelton A, et al. Inhibition of MptpB phosphatase from Mycobacterium tuberculosis 508 impairs mycobacterial survival in macrophages. J Antimicrob Chemother. 2009;63: 509 928–936. doi:10.1093/jac/dkp031 510 29. Dutta NK, He R, Pinn ML, He Y, Burrows F, Zhang Z-Y, et al. Mycobacterial protein 511 tyrosine phosphatases A and B inhibitors augment the bactericidal activity of the 512 standard anti-tuberculosis regimen. ACS Infect Dis. 2016;2: 231–239. 513 doi:10.1021/acsinfecdis.5b00133 514 30. Mechanism of catalysis and inhibition of Mycobacterium tuberculosis SapM, 515 implications for the development of novel antivirulence drugs. Sci Rep. 2019;9:10315. 516 doi.org/10.1038/s41598-019-46731-6 517 31. Rodríguez-Fernández P, Botella L, Cavet JS, Domínguez J, Gutierrez MG, Suckling 518 CJ, et al. MptpB inhibitor improves the action of antibiotics against Mycobacterium 519 tuberculosis and nontuberculous Mycobacterium avium infections. ACS Infect Dis. 520 2024;10: 170–183. doi:10.1021/acsinfecdis.3c00446 521 32. Wong D, Chao JD, Av-Gay Y. Mycobacterium tuberculosis -secreted phosphatases: 522 from pathogenesis to targets for TB drug development. Trends Microbiol. 2013;21: 523 100–109. doi:10.1016/j.tim.2012.09.002 524 33. Mandrekar JN. Receiver operating characteristic curve in diagnostic test assessment. J 525 Thorac Oncol. 2010;5: 1315–1316. doi:10.1097/JTO.0b013e3181ec173d 526 34. Veale JD, Gorman Á, Veale DJ, Fearon U, Orr C, Marzaioli V. Investigation of serum 527 biomarkers in rheumatoid and psoriatic arthritis patients for disease-specific 528 signatures. Arthritis Res Ther. 2025;27: 147. doi:10.1186/s13075-025-03608-6 529 35. Lee D, Schur P. Clinical utility of the anti-CCP assay in patients with rheumatic 530 diseases. Ann Rheum Dis. 2003;62: 870–874. doi:10.1136/ard.62.9.870 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 22 531 36. Naser SA, Thanigachalam S, Dow CT, Collins MT. Exploring the role of 532 Mycobacterium avium subspecies paratuberculosis in the pathogenesis of type 1 533 diabetes mellitus: a pilot study. Gut Pathog. 2013;5: 14. doi:10.1186/1757-4749-5-14 534 37. Sundaresan B, Shirafkan F, Ripperger K, Rattay K, Sundaresan B, Shirafkan F, et al. 535 The role of viral infections in the onset of autoimmune diseases. Viruses. 2023;15: 536 782. doi:10.3390/v15030782 537 38. Rosu V, Ahmed N, Paccagnini D, Gerlach G, Fadda G, Hasnain SE, et al. Specific 538 immunoassays confirm association of Mycobacterium avium Subsp. paratuberculosis 539 with type-1 but not type-2 diabetes mellitus. PLoS One. 2009;4: e4386. 540 doi:10.1371/journal.pone.0004386 541 39. Falkinham JO. Mycobacterial aerosols and respiratory disease. Emerg Infect Dis. 542 2003;9: 763–767. doi:10.3201/eid0907.020415 543 544 Table 1. Clinical and demographic features of RA patients and control subjects (CS). Variable RA (n = 100) Controls (n = 100) P-value Age, years (median, IQR) 54 (41–61) 40 (31–53) 0.90 HAQ score (0–3) 0.75 (0.25–1.25) – – DAS-28 score 3.2 (2.6–5.1) – – C-reactive protein (CRP), mg/dL 6 (2–12) – – Erythrocyte sedimentation rate (ESR), mm/h 20 (12–44) – – White blood cells (WBC), ×10⁹/L 7.0 (5.8–8.5) – – Red blood cells (RBC), ×10¹²/L 4.5 (4.1–4.9) – – Hemoglobin (Hb), g/dL 13.4 (12.1–14.8) – – Mean corpuscular volume (MCV), fL 31.0 (29.0–33.0) – – Platelets (PLT), ×10⁹/L 260 (210–320) – – Rheumatoid factor (RF), IU/mL 45 (20–118) – – Body weight, kg 65 (56–74) – – Height, cm 158 (152–165) – – Body mass index (BMI), kg/m² 27.5 (24.0–31.5) 26.3 (23.6–32 >0.90 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint 23 Disease evolution time, years 8 (2–20) – – Current steroid therapy, n (%) 1 (1%) – – DMARDs, n (%) 73 (73%) – – NSAIDs, n (%) 36 (36%) – – 545 546 Data are expressed as median (interquartile range, IQR) for continuous variables and as absolute numbers with 547 percentages for categorical variables. Age was compared using the Mann–Whitney U test, and sex distribution 548 was analyzed using the chi-square test. A p < 0.05 was considered statistically significant. Abbreviations: RA, 549 rheumatoid arthritis; CS, control subjects; HAQ, Health Assessment Questionnaire; DAS-28, Disease Activity 550 Score 28; DMARDs, disease-modifying antirheumatic drugs; NSAIDs, nonsteroidal anti-inflammatory drugs. 551 552 Table 2. Diagnostic performance of the logistic regression model combining anti-PtpA and 553 anti-PtpB antibody levels. Metric Value Area under the curve (AUC) 0.934 Optimal cut-off point (Youden index) 0.3869 Sensitivity 0.9638 Specificity 0.87 Positive predictive value (PPV) 0.8602 Negative predictive value (NPV) 0.9666 554 .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 8, 2026. ; https://doi.org/10.64898/2026.01.08.698458doi: bioRxiv preprint