Impact of delayed type hypersensitivity arthritis on development of heart failure by aortic constriction in mice.

AIMS: Patients with rheumatoid arthritis (RA) have increased risk of heart failure (HF). The mechanisms and cardiac prerequisites explaining this association remain unresolved. In this study, we sought to determine the potential cardiac impact of an experimental model of RA in mice subjected to HF by constriction of the ascending aorta.
METHODS: Aorta was constricted via thoracotomy and placement of o-rings with inner diameter 0.55 mm or 0.66 mm, or sham operated. RA-like phenotype was instigated by delayed-type hypersensitivity arthritis (DTHA) two weeks after surgery and re-iterated after additional 18 days. Cardiac magnetic resonance imaging (MRI) was performed before surgery and at successive time points throughout the study. Six weeks after surgery the mice were euthanized, blood and tissue were collected, organ weights were documented, and expression levels of cardiac foetal genes were analysed. In a supplemental study, DTHA-mice were euthanized throughout 14 days after induction of arthritis, and blood was analysed for important markers and mediators of RA (SAP, TNF-α and IL-6). In order to put the latter findings into clinical context, the same molecules were analysed in serum from untreated RA patients and compared to healthy controls.
RESULTS: Significant elevations of inflammatory markers were found in both patient- and murine blood. Furthermore, the DTHA model appeared clinically relevant when compared to the inflammatory responses observed in three prespecified RA severity disease states. Two distinct trajectories of cardiac dysfunction and HF development were found using the two o-ring sizes. These differences were consistent by both MRI, organ weights and cardiac foetal gene expression levels. Still, no difference within the HF groups, nor within the sham groups, could be found when DTHA was induced.
CONCLUSION: DTHA mediated systemic inflammation did not cause, nor modify HF caused by aortic constriction. This indicates other prerequisites for RA-induced cardiac dysfunction.

1. Introduction

Both rheumatoid arthritis (RA) and cardiovascular diseases (CVD) have high global prevalence [1, 2]. Approximately 0.5–1.0% of the general population suffers from RA, with considerable variance with respect to hemisphere (more frequent in the northern parts), region (more frequent in urban areas) and sex (more frequent in females) [3], while heart failure (HF) affects approximately 2% of the general population. The risk of developing HF is age-related with a 30% life-time risk of developing HF after 55 years of age [1]. The predominant causes of HF in the western hemisphere are ischemic heart disease and hypertension [4]. However, numerous other conditions, including systemic inflammatory diseases, underlie collectively a substantial proportion of the HF cases [4].

It is well established that RA increases the risk of coronary heart disease [5]. In fact, RA is a numerically defined risk factor in current clinical CVD guidelines [6]. However, it has also been reported that RA increases the risk of HF, even after correction for other CVD risk factors and importantly, exclusion of coronary heart disease [7]. This suggests the presence of RA-specific HF risks. Epidemiological studies suggest an approximately two-fold increased risk of HF with RA [8, 9]. However, these studies do not differentiate between heart failure with reduced ejection fraction (HFrEF) versus heart failure with preserved ejection fraction (HFpEF). Short-duration, prospective studies have shown that an even higher proportion of RA patients develop cardiac dysfunction (i.e., subclinical HF). Of patients with newly diagnosed RA and initial normal heart function, 24% developed mild left ventricular (LV) diastolic dysfunction after one year [10]. However, no longer duration studies have been undertaken, and thus further cardiac deterioration (i.e., into HFrEF) cannot be excluded. Interestingly, the risk of HF is even higher amongst women and appears to increase with age [7]. Finally, data comparing HF patients with or without concomitant RA indicate a doubled mortality in the RA/HF comorbidity as compared to HF alone [11].

Knowledge of the RA-specific HF risk factors is so far limited. RA disease severity has been shown to influence the occurrence of HF [9, 12, 13], and some studies suggest that degree of systemic inflammation is the nodal factor dictating the occurrence and degree of HF [9, 12, 13]. Pharmacological adjustments of RA disease activities also seem to impact HF [12, 14]. However, none of these studies provides definite clues as to whether these findings are due to direct cardiac effects of therapeutic intervention or indirectly through the consequence of decreased systemic inflammation.

Most of the above-mentioned studies are epidemiological. Fewer studies are prospective, and these are all small, of short duration and purely observational. Still, these studies collectively unequivocally document the association between RA and HF. However, they do not allow mechanistic insights, nor address whether RA alone is a sufficient risk factor or require pre-existing cardiac vulnerability. Addressing these highly relevant clinical questions will be exceedingly challenging in human trials and as of yet only a few experimental studies investigating the cardiac impact of experimental RA exists [15-19]. A fundamental question is whether the observed cardiac dysfunction in RA is a consequence of the same pathogenesis (i.e., autoimmunity/autoantibodies) or an indirect effect of resulting systemic inflammation. No studies as of yet have addressed this. However, cardiac dysfunction is also seen in other chronic inflammatory conditions (like inflammatory bowel disease, psoriasis arthritis, ankylosing spondylitis) where the autoimmunogenic pathogenesis is different [20-22]. Furthermore, the few experimental RA models in which cardiac function has been studied also varies in respect to pathogenesis [15-19]. Based on these clinical and experimental observations it is tempting to suggest the common denominator being systemic inflammation.

In this study, we have chosen a different experimental RA model than previously used, the delayed type hypersensitivity arthritis (DTHA) model. Using this model, we have examined its effect on normal hearts, as well as on the natural development of two different HF trajectories. Although not excluded, the regionally restricted arthritis phenotype probably allows for viewing putative cardiac findings as a consequence of arthritis-induced systemic inflammation rather than autoantibodies directly injuring the heart.

2. Methods

2.1 Ethics

The experimental study was performed in accordance with EU directive 2010/63/EU (the Norwegian Food Safety Authority, “Mattilsynet”, approval FOTS ID #14385) and in agreement with the Guide for the Care and Use of Laboratory Animals (NIH publication No. 85–23, revised 2011, US). The reporting of experimental procedures and results were performed in accordance with the ARRIVE guidelines. In the supplemental study we included patients from a clinical study [23] and matched biobanked healthy controls [24]. The clinical study and the biobanked material were in compliance with the Declaration of Helsinki and approved by Regional Committees for Medical Research Ethics in South Eastern Norway (REC South East Norway, reference number 2010–744 [23], 2012–286 and 2016–1540 [24], respectively), and all patients and healthy controls provided written informed consent.

2.2 Mouse o-ring aortic banding (ORAB)

A schematic timeline of the experimental protocol can be viewed in S1 Fig. 10-week-old male C57BL/6J mice (Janvier labs, Le Genest-Saint-Isle, France) weighing 25.0 g (± 0.7) at the time of surgery were housed in a facility with 12-hour light/dark cycle and ad libitum supply of water and standard rodent chow. Mice (n = 50) were housed in cages with 5 in each cage, and were randomized to the ORAB and RA treatments irrespective of the cage. The mice were allowed to acclimatize for at least one week before surgery. Mice were then subjected to thoracic aortic constriction by ORAB surgery with ring inner diameter 0.66 mm (n = 18) or 0.55 mm (n = 18) as previously described [25]. Sham operated mice (mice undergoing the entire surgical procedure except placement of o-ring) were included as controls (n = 14). Decision of degree of ORAB was based upon the above referenced method paper anticipating two distinct trajectories of cardiac hypertrophy/remodelling and dysfunction. In brief, mice were intubated and ventilated with a gas mixture containing 98% oxygen and 2% isoflurane. A skin incision was made on the left side of the thorax, and access to the ascending aorta was established by blunt dissection through the 3rd intercostal space, retraction of the costae and the thymus and removal of periaortic fat and fibrous tissue. A cut o-ring with pre-attached ligatures at each end was subsequently pulled under and around the aorta and tied. Lungs were inflated, incision was covered by the pectoral muscles and the skin closed with continuous sutures. Mice were kept on ventilation until commencing of spontaneous breathing. Mice received analgesia by subcutaneous injection of 0,2 mg/kg buprenorphine pre-operatively and at 8, 16 and 24 hours post-operatively. Additional analgesics were administered subject to animal status.

2.3 Induction and assessment of DTHA in ORAB-operated mice

One week after surgery mice were randomized 1:1 to DTHA or control. Arthritis was induced 14 days after ORAB. In brief, mice were intradermally injected at the base of the tail 2 × 100 μL of methylated bovine serum albumin (mBSA; 2.5 mg/mL in PBS; Sigma-Aldrich, St. Louis, MO, USA), emulsified 1:1 with complete Freund’s adjuvant (CFA; Difco, Detroit, MI, USA) containing 5 mg/mL Mycobacterium butyricum. Four days after intradermal injection mice were intravenously injected 200 μL of anti-collagen II antibodies (5-clone cocktail; 5 mg/mL; Chondrex, Redmond, WA, USA). Finally, after additional three days, arthritis was induced by subcutaneous injection of 20 μL of mBSA (10 mg/mL) in the right hind leg footpad [26]. Sham-RA mice underwent the same injection procedures, but with vehicle. Arthritis phenotype was assessed serially by standardized photography of hind leg (sagittal projection) towards a ruler. Photographs were subsequently used to measure hind leg diameter using ImageJ. As the DTHA model is of transient nature, arthritis was re-established by a subsequent i.v. injection of anti-collagen II antibodies at day 15 after initial induction of arthritis, and subsequent s.c. injection of mBSA at day 18.

2.4 Cardiac magnetic resonance imaging

Cardiac magnetic resonance imaging (MRI) was performed on 7 mice from each of the six intervention groups (ORAB 0.55 and 0.66, and sham, with or without RA) at day 14, 28 and 42 after ORAB surgery. Mice were randomly picked from each group for MRI scanning before entering the arthritis-inducing protocol. In addition, a randomized selection of mice (n = 12) were examined by MRI one week before ORAB surgery in order to generate an unpaired baseline.

MRI scanning was performed as previously described [25, 27]. In brief, the 9.4 T MRI system (Agilent, Palo Alto, CA, USA) with a 35 mm quadrature-driven birdcage RF coil (Rapid Biomedical, Rimpar, Germany) dedicated to mouse imaging was used. Anaesthesia given by mask during the MRI examination was 1–2.5% isoflurane and O2. Mice were positioned prone in a dedicated animal bed. Body temperature, respiration and ECG were continuously monitored during examination. Temperature was automatically maintained using heated air. The level of anaesthesia was manually adjusted throughout the examination to ensure stable respiration and heart rate. Cine-MRI was acquired in using a stack of short-axis slices covering the LV, and TPM-MRI was acquired in four-chamber long-axis orientation. All MRI data was ECG- and respiratory-gated. Central imaging parameters were repetition time 5.0 ms for cine-MRI and 4.48 ms for tissue phase mapping (TPM)-MRI, acquisition matrix 128 x 128 and 96 x 96 for cine- and TPM-MRI, respectively. Field-of-view was 25 x 25 mm and slice thickness 1 mm for both cine- and TPM-MRI.

2.4.1 MRI post-processing

Segmentation of the cine-MRI was conducted using Matlab (The MathWorks Inc., USA), where endocardium and epicardium were detected in a semiautomatic manner. Briefly, the semiautomatic method took the centrum of the left ventricle as user input and converted the images into polar coordinates. The contrast between blood and myocardium were automatically detected and were used to generate endo- and epicardial masks. These masks were then manually validated and adjusted, if necessary, by a single examiner. From the cine-MRI data, left ventricular (LV) mass, LV end-diastolic volume (LVEDV), LV end-systolic volume (LVESV), stroke volume (SV), and LV ejection fraction (LVEF) were calculated.

Analysis of the TPM-MRI data was conducted using Matlab (The MathWorks Inc., USA) as previously described in detail [27]. TPM-MRI was used to measure global longitudinal strain (GLS, %), peak early diastolic strain rate (SRe’, 1/s) and peak systolic strain rate (SRs’, 1/s) in the septum and the LV free wall, and global. SRs’ and SRe’ were measured from the peaks in the strain rate curve gained from TPM using Matlab (The MathWorks Inc., USA). All MRI-analysis was conducted by staff blind to the treatment of the animals.

2.5 Tissue- and blood sampling

Mice were euthanized deeply anesthetized in a mixture of oxygen and 3–4% isoflurane by extraction of the heart. Organs were harvested and weighed before the heart were cut at the transverse midventricular plane. The apical part of the LV was snap frozen in liquid nitrogen and stored at -80°C until use. In the supplemental study, mice were euthanized by cutting the carotid artery as described in 2.7.

2.6 Isolation of RNA and analysis of myocardial gene expression

Total RNA was isolated from myocardial tissue samples by homogenization in chaotropic salts and subsequent ion exchange chromatography with the RNeasy system (Qiagen) and kept in RNase-free water containing RNase inhibitor (RNasin; Promega) at −70°C until further use. For real-time qPCR, total RNA was reverse transcribed with the TaqMan Reverse Transcription Reagents Kit, and subsequently real-time qPCR of each sample was run in triplicate with TaqMan Pre-Developed Assay Reagents and the StepOnePlus™ Real-Time PCR System and software (Applied Biosystems, Foster City, CA) according to the manufacturer’s instructions. A standard curve was obtained by amplifications of cDNA obtained from serial dilutions of myocardial RNA. Atrial natriuretic peptide (NPPA), brain natriuretic peptide (NPPB), β-myosin heavy chain (MYH7), and skeletal alpha actin (ACTA1) were measured. For all specific mRNA amplified, linear inverse correlations were observed between amount of mRNA and CT value (number of cycles at threshold lines). Gene expression is presented relative to the levels of 18S rRNA as the housekeeping gene.

2.7 Assessment of degree of clinically relevant systemic inflammatory responses in the DTHA model

To validate the clinical translational value of the inflammatory responses in the circulation of the murine DTHA model, we conducted a supplemental time course study where mice subjected to DTHA or control protocol were euthanized sequentially for 14 days. Mice were anesthetized with a gas mixture containing 96–97% oxygen and 3–4% isoflurane, and a midline incision was made from the mandibulae towards the sternum. With blunt dissection the carotid artery was identified and subsequently cut. 50 μL 0.5 M EDTA was deposed within the incision followed by extraction of blood into a 100cc syringe. Blood was transferred to a prechilled 1.5 mL centrifuge tube and subsequently centrifuged at 4°C for 20 minutes at 2000 × g. Plasma was aliquoted and snap frozen in liquid nitrogen and stored at -80°C until use. Mouse plasma was subsequently analysed for SAP using enzyme immunoassay (EIA) (Hycult Biotech, Uden, Netherlands), and IL-6 and TNF-α were analysed using MSD cytokine assay (Meso Scale Discovery, MSD, Rockville, MD, USA). To validate the significance of the relative increments of key inflammatory parameters in the DTHA model we analysed CRP, IL-6 and TNF-α in plasma from patients with untreated seropositive RA. The patient characteristics have previously been published [23]. This study, however, did not include age- and sex-matched healthy volunteers, a necessary prerequisite for assessing the relative increments of inflammatory mediators needed for comparing the human and murine data. Such controls were recruited by the Norwegian PSC Research Center biobank after providing written informed consent [24]. Patients were drawn from the ARCTIC biobank according to disease severity assessed by Disease Activity Score (DAS), and thus categorized in a mild, moderate and severe disease group. Human plasma was analysed using MSD cytokine assay for IL-6 and TNF-α, and EIA was used for CRP (RnD systems, Stillwater, MN). Intraassay coefficients of variation were <10%.

2.8 Statistics

Data and figures are presented as means with standard error of the mean. The statistical calculations were made in GraphPad Prism 9 statistical package (GraphPad Software, La Jolla, CA). For comparisons of two groups, we used unpaired t-test. For non-repeated measures with more than two groups, we used one-way ANOVA with Bonferroni’s multiple comparisons test. For repeated measures, we used two-way ANOVA with Bonferroni’s multiple comparisons test. To identify outliers, we used the ROUT method with a Q value of 1%. Statistical significance was defined as P < 0.05.

2.9 Evaluation of animal status

We evaluated the animal status daily with a score sheet with prespecified criteria, based on posture, fur, grooming, weight loss, and clinical features of limited gait. Based on the score we assessed the need for pain relief and closer monitoring, or if euthanization was necessary. In addition, a footpad swelling > 4 mm compared to baseline or a wound at injection site > 8 mm required immediate euthanization.

2.10 Calculating group sizes

Group sizes were estimated using power analysis. We chose a significance level of 0.05 and power of 0.8 resulting in constant C = 7.85 according to Guidelines for the Care and Use of Mammals in Neuroscience and Behavioural Research [28]. Based on current available data [15], we hypothesised a mean difference (D) in heart function of 15% between RA+ and RA-, measured by ejection fraction (EF, %) on MRI, and a standard deviation (SD) of 10%. The number of required animals is given by:

n = 1+2C (SD/D)2. Hence n = 1+15.7(0.10/0.15)2 = 8. As we expected a slightly higher standard deviation in the ORAB-groups than in sham animals, we chose to have only 7 animals in the sham groups, and more animals in the ORAB-groups.

2.11 Criteria for exclusion of animals

We did have preset criteria for exclusion of animals according to failure of surgery, faulty induction of arthritis, animal welfare (see Methods section 2.9 above) and so forth. However, none of these were met (i.e., none were excluded).

2.12 Blinding

TCT was the only investigator aware of the group allocation during the conduct of the experiment, and blinding was initiated before experimentation and analysis involving TCT.

3. Results

3.1 Inflammatory responses in the DTHA model

3.1.1 Establishment and assessment of the DTHA model

Arthritis phenotype was seen in 100% of the animals in accordance with the original method paper [26]. The phenotype was macroscopically restricted to the right hind leg (where the antigen was s.c. injected) and the variance was low (see Fig 1, panel A). Significant swelling was observed from day 1 through day 8. Also, as seen in Fig 2, panel D-F, we observed a robust, although transient increase of clinically relevant inflammatory mediators in circulation. Figures displaying absolute data are shown in S2 Fig, panel D-F. One mouse subjected to the ORAB 0.66 DTHA group died at day 12 after establishment of DTHA for unknown reasons. There were no other indications that mice subjected to DTHA failed to thrive as they had a similar gain of body weight as compared to the control mice (Fig 3A).

Fig 1

Assessment of footpad swelling in the DTHA model.

Panel A: Footpad swelling from day 0 through day 28 after arthritis induction by s.c. injection of mBSA (RA1). At day 15, mice received an additional i.v. injection of anti-collagen II antibodies before an additional injection with mBSA (day 18, RA2), re-establishing arthritis. Figure displays increase of sagittal diameter in mm from baseline. Data is shown as mean ± SEM (n = 25). mBSA, methylated bovine serum albumin. Panel B: Representative photographs displaying temporal paw swelling.

Fig 2

Analysis and comparison of systemic inflammation in the DTHA model and RA patients.

Panel A-C: Serum levels of CRP (panel A), IL-6 (panel B) and TNF-α (panel C) from RA patients. Patients were chosen according to the DAS to represent three disease-states of severity (mild, moderate and severe). Each group consisted of 12 patients and age- and sex-matched healthy volunteers (control). CRP was analysed individually, and included 12 individuals in each group. One outlier was excluded in the healthy volunteer group. Due to space limiting reasons (limited wells in the 96-well) in the analysis of IL-6 and TNF-α, 10 patients were randomized for analysis in each disease group. Of these we excluded one outlier of IL-6 in the moderate group, one outlier of TNF-α in the control group. In addition, we excluded one IL-6 value in the control group, even though it did not meet our exclusion criterion, due to that the same individual had outlying both CRP and TNF-α values. Data is displayed as relative to control and presented as scatter plots with mean ± SEM. Statistical significance was calculated using one-way ANOVA with Bonferroni’s multiple comparisons test for CRP and IL-6, and unpaired two-tailed t-test of control vs intervention (mild, moderate and severe combined) for TNF-α. Panel D-F: Temporal profiles of serum levels of murine SAP (panel D), IL-6 (panel E) and TNF-α (panel F) in DTHA mice relative to control. n = 10 for control group, n = 5 for other time points. Of these, two outliers for both SAP and IL-6 (representing the same two mice) were excluded in the control groups. Even though it did not meet our exclusion criterion, we chose to also exclude the corresponding TNF-α values from these two mice. Red grid line in graphs represents corresponding mean relative human value in severe human RA patients, blue grid line in graphs represents corresponding mean relative human value in moderate human RA patients, green grid line in graphs represents corresponding mean relative human value in mild human RA patients. Data is shown as mean ± SEM. S = severe, Mo = moderate, Mi = mild. Statistical significance was calculated in mice, relative to control, using one-way ANOVA with Bonferroni’s multiple comparisons test. ns = non-significant, *P < 0.05, †P < 0.01; ‡P < 0.001, #P < 0.0001. DTHA, delayed type hypersensitivity arthritis; CRP, C-reactive protein; IL-6, interleukin 6; TNF-α, tumor necrosis factor α; DAS, Disease Activity Score; SAP, serum amyloid P component.

Fig 3

Temporal body weight alterations and organ weights.

Panel A: Development of mice body weight within the various intervention groups throughout the timespan of the study (n = 7, 7, 8, 9, 10, 8 for sham RA-, sham RA+, ORAB 0.66 RA-, ORAB 0.66 RA+, ORAB 0.55 RA- and ORAB 0.55 RA+, respectively). Day 0 represents data before ORAB surgery, arthritis was induced at 2 weeks (RA1). At day 18 after RA induction, arthritis was re-established with an additional injection with mBSA (RA2). Data is mean ± SEM. Panel B-D: Total heart weight, left ventricular weight, and lung wet weight at the time of euthanization (n = 7, 7, 8, 9, 10, 8 for sham RA-, sham RA+, ORAB 0.66 RA-, ORAB 0.66 RA+, ORAB 0.55 RA- and ORAB 0.55 RA+, respectively). Data is presented as scatter plots with mean ± SEM. Statistical significance was calculated using one-way ANOVA with Bonferroni’s multiple comparisons test. ns = non-significant, *P < 0.05, †P < 0.01; ‡P < 0.001, #P < 0.0001. ORAB, o-ring aortic banding; mBSA, methylated bovine serum albumin.

3.1.2 Validation of systemic inflammatory responses and the putative clinical relevance of the DTHA model

As this study aimed to experimentally address the putative clinical negative association between RA and cardiac function, establishing the clinical relevance of our RA model was highly important. Clinical relevance in this aspect reflects the model’s pathophysiological relevance, type of immune response as well as the magnitude and duration of the inflammatory response. As previously described, the DTHA model displays a pathogenesis at least partly resembling the human counterpart [26, 29]. The experimental arthritis phenotype is triggered by an unspecific antigen (mBSA, this resembling a neoantigen) and it generates a familiar local inflammation with systemic “spillover”. Furthermore, the DTHA-induced arthritis can be abrogated by the most common human anti-inflammatory biologic drugs [29]. Although not studied, the DTHA model does probably not convey “autoantibody”-mediated actions on other tissues than the neoantigen-injected hind-limb. Thus, the DTHA model may represent a more isolated aspect (systemic inflammation) of RA-mediated cardiac effects, an aspect which is not as easily deciphered in other experimental RA-models.

We determined the relative alterations of important inflammatory markers and mediators of RA in the DTHA model. As seen in Fig 2, panel D-F, both SAP (the murine counterpart of CRP), IL-6 and TNF-α were significantly elevated after induction of arthritis. To assess the clinical relevance of these alterations in the DTHA model, we analysed the human counterparts of these inflammatory mediators in serum from untreated RA patients and healthy controls. The details regarding patient characteristics have previously been published [23, 24]. We prespecified three groups of disease severity according to the DAS and within these groups we randomly selected patients for further analysis, keeping the male-to-female ratio at 1. Patient demographic data is shown in S1 Table. Fig 2 shows the increases of CRP (panel A), IL-6 (panel B) and TNF-α (panel C) in the three disease severity groups, relative to control. Graphs displaying absolute values are shown in S2 Fig. As expected, when analysing a relatively small number of patient samples, variance is high, but with a clear trend for increased values for all measured markers. For IL-6 levels, all the three disease severity groups were significantly higher than control. Furthermore, when combining the three disease severity groups, levels of TNF-α and CRP were also significantly higher than the controls. Between the disease severity groups, only CRP and IL-6 were significantly different between mild and severe human RA. We did not statistically compare the human and mice results, as this would be futile due to the large variance in the data. However, we have graphically displayed the human average values (mild disease: green line, moderate disease: blue line and severe disease: red line) together with the mouse data (in Fig 2, panel D, E and F). Although not statistically compared, we believe the data overall supports the notion that our experimental DTHA model gives a clinically relevant humoral inflammation within the time span of ORAB-induced HF development.

3.2 HF parameters in the ORAB model

Cardiac dysfunction and heart failure was induced by constriction of the ascending thoracic aorta by a modification of previously established methodology. This refined technique allowed us to generate highly predictable cardiac phenotypes with low variance, as described in the method section [25]. Two different o-ring diameters were chosen to constrict the ascending aorta leading to cardiac stress, causing two different trajectories of cardiac dysfunction and HF [25]. There was highly significant step-wise increases in total heart weight and LV weight between sham-operated mice, ORAB 0.66 and ORAB 0.55 (Fig 3, panel B and C). These findings were corroborated by MRI where estimated LV mass for ORAB 0.55 and ORAB 0.66 were significantly higher than sham-operated mice at all post-surgery time points studied (P < 0.0001 for all comparisons made). Furthermore, the severity of aortic constriction afforded by ORAB 0.66 and ORAB 0.55 generated two distinct trajectories of cardiac hypertrophy and cardiac dysfunction as shown in Fig 4A and 4B. More interestingly, the two chosen degrees of aortic constriction yielded a binary result as to pulmonary congestion (Fig 3, panel D), as only ORAB 0.55 displayed increased lung weight.

Fig 4

Analysis of LV ejection fraction and LV myocardial mass.

MRI assessed LV ejection fraction (EF %; panel A) and LV myocardial mass (panel B) at given time points (n = 42 in total, 7 in each group). Data is presented as mean ± SEM. RA1 = time of establishment of RA phenotype with a sc. injection of mBSA, RA2 = time of re-establishment of RA phenotype with an additional injection with mBSA. Statistical significance was calculated using two-way ANOVA with Bonferroni’s multiple comparisons test. *P < 0.05, †P < 0.01; ‡P < 0.001, #P < 0.0001. MRI, magnetic resonance imaging; LV, left ventricular; mBSA, methylated bovine serum albumin.

3.3 Assessment of cardiac function after ORAB surgery

There was a steady, significant decline in systolic and diastolic function over time for both ORAB 0.55 and ORAB 0.66, as compared to sham-operated mice. These data were substantiated by measurement of LVEF (Fig 4, panel A), GLS (%, Fig 5, panel A), peak systolic strain rate (SRs’, Fig 5, panel C), and peak early diastolic strain rate (SRe’, Fig 5, panel B). All these measurements were showing a significant difference between sham and the two ORAB groups, with P <0.0001 for all comparisons made. Between the ORAB groups there was a significant difference in LVEF, SRs’ and GLS between 0.66 vs 0.55 from the first MRI at week 2 (P = 0.015 for LVEF, P < 0.0001 for SRs’ and GLS, respectively). These observations were substantiated by findings of mirroring foetal gene alterations (Fig 6, panel A-D). All four foetal gene expression levels were significantly elevated after ORAB surgery as compared to sham, but the data also substantiate the evidence for the step-like difference in HF development previously described, as mRNA levels of NPPA and MYH7 were significantly higher in ORAB 0.55 compared to ORAB 0.66 (Fig 6, panel A and C).

Fig 5

Analysis of heart function parameters by TPM-MRI.

MRI assessed LV strain (%, panel A), strain rate e’ (panel B; SRe’, peak early diastolic strain rate) and strain rate s’ (panel C; SRs’, peak systolic strain rate) in septum and LV free wall combined at given time points (n = 42 in total, 7 in each group). Values are mean ± SEM. RA1 = time of establishment of RA phenotype with a sc. injection of mBSA, RA2 = time of reestablishment of RA with an additional injection with mBSA. Statistical significance was calculated using two-way ANOVA with Bonferroni’s multiple comparisons test. ns = non-significant, *P < 0.05, †P < 0.01; ‡P < 0.001, #P < 0.0001. MRI, magnetic resonance imaging; LV, left ventricular; mBSA, methylated bovine serum albumin.

Fig 6

Foetal gene activation in HF with or without arthritis.

Myocardial gene expression levels of NPPA (panel A), NPPB (panel B), MYH7 (panel C) and ACTA1 (panel D) was analysed at termination of the experimental study, relative to the housekeeping gene 18S (n = 7, 7, 8, 9, 10, 8 for sham RA-, sham RA+, ORAB 0.66 RA-, ORAB 0.66 RA+, ORAB 0.55 RA- and ORAB 0.55 RA+, respectively). One outlier was excluded in the analysis of ACTA1 in the 0,66 RA- group. Data is displayed as scatter plots with mean ± SEM. Statistical significance was calculated using one-way ANOVA with Bonferroni’s multiple comparisons test. ns = non-significant, *P < 0.05, †P < 0.01; ‡P < 0.001, #P < 0.0001. NPPA, natriuretic peptide A; NPPB, natriuretic peptide B; MYH7, myosin heavy chain 7; ACTA1, actin alpha 1, skeletal muscle.

3.4 Impact of arthritis on HF development

As previously described, we believe our model of RA represents relevant systemic inflammation in the context of HF. Furthermore, we reproduce that the ORAB model yields HF with graded clinical cardiac trajectories with clear evidence of both systolic and diastolic dysfunction. Still, when combining the models, no overall difference in HF development could be seen between mice with and without arthritis. The only numerical significant difference between arthritis and control was seen in a few isolated data sets in the ORAB 0.66 group (MRI assessed LV myocardial mass, SRe’ and total body weight). Although this could leave the impression of an inflammatory beneficial effect on mild HF, it is important to emphasize that these differences appeared before induction of DTHA. Furthermore, similar differences could not be found in the more severe ORAB group, nor in the sham group. Thus, we conclude that these differences most likely is not due to arthritis. All the other data including MRI (Figs 4 and 5), foetal gene expression levels (Fig 6) and organ weights (Fig 3) provide in sum multiple lines of evidence that in this setting, the chosen experimental model of RA does not impact the development of HF in mice.

4. Discussion

RA is associated with HF [7, 9, 12, 30, 31] and data suggests there are RA-specific HF risk factors [7]. These risk factors correlate to RA disease activity [9] and are probably mediated through unknown aspects of systemic inflammation. Currently, it is unclear whether isolated RA is a sufficient promoter of HF, or whether RA needs to act in concert with an already HF-prone heart in order for HF to develop. Thus, the mechanisms underlying the association between RA and HF are unknown. We believe gaining such in depth insights will be challenging in human trials. On this note, although imperfect, experimental animal studies may provide data of potential clinical relevance on these questions. Such knowledge will be important when choosing RA treatments and in the development of specific treatments for HF within the RA/HF comorbidity, as well as identifying RA patients at risk of developing HF.

The main finding of the current study is that DTHA-mediated arthritis did not cause, nor influence the evolvement of HF in the ORAB HF model. We temporally assessed alterations of cardiac dimensions and function, alterations of organ weights and activation of the cardiac foetal gene program. All displayed clear differences between the two HF groups, as well as between the HF groups and sham control, but were not influenced by the presence of arthritis. Cardiac dysfunction/HF in RA has convincingly been demonstrated clinically, as well as in three different mouse RA models [15, 17, 18] and one rat RA model [16, 19]. Thus, the pivotal question in the current study is why we did not observe the same. There are several experimental parameters in the current study that could explain why DTHA-mediated arthritis did not impact cardiac function, ranging from type, degree and duration of systemic inflammation, type and duration of HF, sequence of disease introduction (HF vs arthritis), as well as mouse strain and sex.

Firstly, we believe the current study’s experimental models of HF and RA are relevant of human disease [25, 26], and that combining these models is a valid way of observing the impact of arthritis-induced systemic inflammation on the natural evolvement of HF. By standardised constriction of the ascending aorta by different sized o-rings, cardiac dysfunction/HF was induced with a phenotype resembling in mechanism that of aortic stenosis and hypertension. It is important to emphasize that this HF model produces predictable cardiac pathological trajectories with a very low variation. As is clear in the current study, our chosen degrees of aortic constriction generated distinct paths of HF development, with clear and evolving cardiovascular dysfunction, but importantly with near binary outcomes in resulting cardiac decompensation and backward failure.

No ideal experimental model of RA exists, as human RA is considered a heterogenous disease with varying aetiology [32, 33] and there are major knowledge gaps regarding the pathogenesis [11, 14]. Common traits are considered to be the generation of neoantigens, which in turn activates both the innate and adaptive immune system and subsequent generation of arthrogenic autoantibodies [34, 35]. In such, the DTHA model mimics some aspects of the RA pathogenesis, as mBSA may represent the neoantigen and injecting anti-collagen II antibodies may mimic the generation of autoantibodies. Previous investigations have also demonstrated that arthritis in the DTHA model is treatable with clinically used biological disease modifying anti rheumatic drugs [29]. However, and as previously commented on, we believe that any systemic effect of DTHA-mediated local arthritis would not be a consequence of autoimmunity, this because the arthritis phenotype observed is exclusively restricted to the mBSA-injected hind limb. Therefore, and of putative importance for interpretating the current study, any observed cardiac effects in the DTHA model would most likely be a consequence of systemic “spillover” of locally generated inflammation.

Thus, the most important aspect of the clinical validity of the current study, we believe to be the degree and duration of systemic inflammation in the DTHA model. To quantify systemic inflammation, we conducted a time study of DTHA and analysed levels of important markers and mediators of inflammation. To place these experimental data in clinical context, we analysed serum from untreated RA patients categorized with mild, moderate or severe disease according to a globally endorsed Disease Activity Score (DAS) and from healthy controls. Variation in the two data sets (murine vs human), and especially in the patient material, did not allow for adequate statistical comparison between murine and patient inflammatory responses. Such comparison could have provided some cues as to the clinical value of the DTHA model (i.e., the time-amount mice with HF were exposed to “clinical relevant inflammation”). However, as should be evident when studying the data as a whole, the DTHA model does indeed elicit systemic inflammation that both spatially and temporally appears clinically relevant, but albeit may still be of inadequate strength or duration.

This is the first experimental study where putative cardiac effects of DTHA-induced arthritis has been studied. Although few, some previous investigations on both mice and rats using three different models of experimental RA have demonstrated cardiac dysfunction [15-19]. Both diastolic dysfunction [16, 17, 19] and systolic dysfunction [15, 18] have been demonstrated. These studies share some common features that may explain the apparent difference in findings, i.e., their observed negative cardiac effects towards our lack of such. One obvious difference is that these studies, in contrast to our study, all display systemic, antibody-mediated polyarthritis. This could imply a common pathogenesis of arthritis and cardiac dysfunction, this being autoantibody-mediated myocardial damage. Although not excluded, we believe such a hypothesis is less likely as the antibodies involved in these experimental RA-models are very different in generation and action. Clinical data showing an association between very different autoimmune diseases and HF also support the notion of another common denominator than autoantibodies [20-22].

As to our knowledge, none of the above studies have examined the duration and magnitude of systemic inflammation elicited by their RA-models. As argued for above, we believe the DTHA-induced systemic inflammation to be clinically relevant. Even so, the other RA-models may elicit a more robust and durable systemic inflammation due to their polyarthritic nature, as well as visually being of more chronic nature. In fact, Mokotedi et al did find a correlation between levels of inflammatory mediators and degree of cardiac dysfunction in the collagen-induced arthritis rat model [16]. The latter harmonizing with clinical observations of a correlation between RA disease degree and cardiac dysfunction [9, 12, 13]. We ended our study at day 28 after induction of arthritis and some of the other studies have had longer observational periods [15-17]. Thus, even though we believe it unlikely, we cannot exclude that extending our observation period would reveal differences in cardiac dysfunction.

Finally, the risk of cardiac dysfunction/HF amongst RA patients probably requires a genetic predisposition. In studies where cardiac function in experimental mouse RA-models has been studied, the strain has not been C57BL/6J (which was used in the current study). Thus, we cannot exclude the possibility that such genetic differences, in this case C57BL/6J vs BALBc or DBA/Ij, may play a role in the susceptibility for cardiac dysfunction in systemic inflammation.

To conclude, even though we believe that the DTHA model is a clinically relevant model of systemic inflammation in RA, it did not induce cardiac dysfunction, nor alter the course of HF induced by aortic constriction, within our set experimental boundaries. In context with the other experimental studies that have found cardiac dysfunction in RA, the current study provides a foundation for further studies exploring the requirements needed for RA to induce cardiac dysfunction/HF.

Graphical abstract.

Schematic timeline of the experimental procedures.

(TIF)

Click here for additional data file.

Analysis and comparison of systemic inflammation in the DTHA model and RA patients (absolute values).

Panel A-C: Serum levels of CRP (panel A), IL-6 (panel B) and TNF-α (panel C) from RA patients. Patients were chosen according to the DAS to represent three disease-states of severity (mild, moderate and severe). Each group consisted of 12 patients and age- and sex-matched healthy volunteers (control). CRP was analysed individually, and included 12 individuals in each group. One outlier was excluded in the healthy volunteer group. Due to space limiting reasons (limited wells in the 96-well) in the analysis of IL-6 and TNF-α, 10 patients were randomized for analysis in each disease group. Of these we excluded one outlier of IL-6 in the moderate group, one outlier of TNF-α in the control group, and one outlier of CRP in the control group. In addition, we excluded one IL-6 value in the control group, even though it did not meet our exclusion criterion, due to that the same individual had outlying both CRP and TNF-α values. Data is displayed as absolute values and presented as scatter plots with mean ± SEM. Statistical significance was calculated using one-way ANOVA with Bonferroni’s multiple comparisons test for CRP and IL-6, and unpaired two-tailed t-test of control vs intervention (mild, moderate and severe combined) for TNF-α. Panel D-F: Temporal profiles of serum levels of murine SAP (panel D), IL-6 (panel E) and TNF-α (panel F) in DTHA mice. Absolute values are shown. n = 10 for control group, n = 5 for other time points. Of these, two outliers for both SAP and IL-6 (representing the same two mice) were excluded in the control groups. Even though it did not meet our exclusion criterion, we chose to also exclude the corresponding TNF-α values from these two mice. Data is shown as scatter plots with mean ± SEM. Statistical significance was calculated using one-way ANOVA with Bonferroni’s multiple comparisons test, relative to control. ns = non-significant, *P < 0.05, †P < 0.01; ‡P < 0.001, #P < 0.0001. DTHA, delayed type hypersensitivity arthritis; CRP, C-reactive protein; IL-6, interleukin 6; TNF-α, tumor necrosis factor α; DAS, Disease Activity Score; SAP, serum amyloid P component.

(TIF)

Click here for additional data file.

Baseline characteristics.

Baseline characteristics of patients and healthy controls in the supplemental study. Data is presented as mean values with 95% confidence interval. DAS, Disease Activity Score; SJC-44, swollen joint counts among 44 joints; CRP, C-reactive protein; SR, sedimentation rate.

(TIF)

Click here for additional data file.

Dataset.

(XLSX)

Click here for additional data file.

8 Oct 2021

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Michael Bader

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please modify the title to ensure that it is meeting PLOS’ guidelines (https://journals.plos.org/plosone/s/submission-guidelines#loc-title). In particular, the title should be "specific, descriptive, concise, and comprehensible to readers outside the field" and in this case we feel it is not informative and specific about your study's scope and methodology.

3. We note that you have stated that you will provide repository information for your data at acceptance. Should your manuscript be accepted for publication, we will hold it until you provide the relevant accession numbers or DOIs necessary to access your data. If you wish to make changes to your Data Availability statement, please describe these changes in your cover letter and we will update your Data Availability statement to reflect the information you provide.

4. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This manuscript reports an interesting study aiming to address a highly relevant topic; the effect of RA on the development of heart failure. Unfortunately, there are a few concerns and publication at this stage seems premature.

Major comments:

+ Although not many, there are studies on cardiac effects in animal models of RA, especially CIA (e.g. Reynolds et al, 2012; He et al 2013; Palma et al 2016, reviewed in Sanghera et al 2019), which should be discussed.

+ Structure/reporting: The value of this studies clearly lies in the fact that this model of DTH-RA has been tested in conjunction with HF. There is obvious value in reporting the findings for other researchers who might want to test a similar hypothesis, but the study should then be described accordingly with an emphasis on the models, rather than overselling relevance and impact of a potential difference that is not present. The manuscript in its present form focuses too much on a potential difference in outcome +/-RA which has not been observed.

+ The choice of experimental design is surprising. Systemic autoimmune diseases underly cardiac complications and may exacerbate development of heart disease. While this is well established for a range of autoimmune conditions, the clinical relevance of pressure-induced HF followed by autoimmunity is questionable. If the authors aim to address the clinical problem they refer to, autoimmunity needs to be induced before heart failure not after.

+ The model used here is delayed-type hypersensitivity (which is not autoimmunity) with auto-Abs targeting an individual antigen (collagen). While no model of systemic autoimmunity is perfect, there are well accepted RA models that reflect human pathology better. Some of them can be time consuming and rather tedious to induce, so it is much appreciated that alternatives are being sought, but this needs to be reflected in the description of the study, starting with a precise title.

Minor comments:

+ It’s unclear why well established patient data should be repeated in a suboptimal sample size. Why not reference previous reports?

+ There are a few instances where ‘outliers have been excluded’. It is unclear what criteria were used to define outliers and how exclusion has been justified.

+ There appears to be a small beneficial effect or the DTH-RA model on mild HF in some of the readouts? This might warrant discussion and further investigation considering that ‘fresh’ inflammation might in fact be beneficial if introduced at the right time.

Reviewer #2: The authors have sought to investigate the link between RA and heart failure by assessing the impact of RA on cardiac function using a murine model of thoracic aortic constriction. As they indicate, RA patients are at increased risk of HF and the mechanisms driving this are still to be defined. This is therefore an important study and the manuscript is well written.

Comments:

1. It is interesting that the authors have chosen to study the effect of arthritis on the development of HF rather than the effect of arthritis on the development of cardiac dysfunction, although the rationale for this is explained. They have chosen an aortic constriction model driven by increased LV afterload. In the introduction both HFpEF and HFrEF is discussed and the literature suggests that RA patients may be particularly susceptible to HFpEF, which is driven by systemic inflammation (PMID: 34523821). However, the model chosen by the authors results in impaired EF. The authors are correct that there are few animal models to investigate HF in the context of RA and indeed models of diastolic dysfunction or HFpEF although there a few more examples in the literature that the authors may wish to discuss as the Pironti study is not the only experimental study that has investigated the cardiac consequence of RA as claimed in the discussion (PMID: 32208438; PMID: 33427610; PMID: 34526398). The authors might want to add a few lines to the discussion to distinguish between the different models of cardiac dysfunction and their relevance to the RA community.

2. A timeline for the experimental protocol would be useful indicating when ORAB was initiated and then when RA was induced.

3. Figure 1 panel A is this data mean +/- SEM (n=25) as indicated for panel B? Please add to panel A legend. No need to indicate ORAB in this legend as this data is not shown here.

4. Figure 2 (panel A-C) I am not sure what the authors mean by ‘due to space limiting reasons in the analysis……, 10 patients were randomised for analysis in each disease group’, yet there are 11 data points per group and it states that one outlier has been removed from each group. It is also not clear to me looking at the absolute values in supplemental figure 1, which patient(s) have been excluded as outliers as well. Also, why show the data as relative to control and then have another figure with the absolute values, rather than just showing absolute values in the main manuscript. There is some variance, as the authors have stated, but this is to be expected with human samples.

5. Were cytokine levels assessed in mice after the re-establishment of RA at days 15/18?

6. Panel 2 (D-F) as above, I am not sure why 2 data points have been excluded for SAP and IL-6 and not TNF and again why data is shown as relative to control as well as absolute data in the supplemental file. Is this because statistical significance was not reached in the murine data when using absolute values? It states in the text, lines 303-306 that ‘Although non-significant due to variance in the data, we believe the data overall supports the notion that our experimental DTHA model…’ This leads me to believe that the absolute data shown in supplemental figure 2 is being referred to here, not the relative data shown in figure 2 where significance is shown. I agree with the author’s interpretation of the data and would just show absolute values as shown in the supplemental figure. I do not think you can correlate the relative data in mice to levels in patients, or that it is necessary to do so.

7. Line 303: It is not clear here that the text is now describing the murine data. Line 304 states CRP, rather than SAP.

8. Lines 407-411 in the discussion are not clear to me. I am not sure what the authors mean by time-amount mice with HF were exposed to ‘clinical relevant inflammation’.

Minor comment

Yellow lines are difficult to see clearly in the figures.

Legend for figure 3 panel B-D ends ‘mBSA, methylated’. Sentence is incomplete, but mBSA is not shown in the figure so this can be deleted – also in the legends for figures 4 and 5.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

2 Dec 2021

Reviewers' comments:

Reviewer #1: This manuscript reports an interesting study aiming to address a highly relevant topic; the effect of RA on the development of heart failure. Unfortunately, there are a few concerns and publication at this stage seems premature.

Major comments:

+ Although not many, there are studies on cardiac effects in animal models of RA, especially CIA (e.g. Reynolds et al, 2012; He et al 2013; Palma et al 2016, reviewed in Sanghera et al 2019), which should be discussed.

+ Structure/reporting: The value of this studies clearly lies in the fact that this model of DTH-RA has been tested in conjunction with HF. There is obvious value in reporting the findings for other researchers who might want to test a similar hypothesis, but the study should then be described accordingly with an emphasis on the models, rather than overselling relevance and impact of a potential difference that is not present. The manuscript in its present form focuses too much on a potential difference in outcome +/-RA which has not been observed.

Answer: We do appreciate these comments, comments also raised by reviewer #2. We believe that addressing these comments have significantly elevated the quality of the paper. The findings in these studies against the (lack of) findings in the current study have now been discussed in detail, with emphasis on putative explanations for the different results. The current manuscript has been thoroughly revised positioning our findings in context of previous investigations, and in particular problematized different aspects of our choice of experimental models and setup. In such, we hope the current manuscript appear more nuanced.

+ The choice of experimental design is surprising. Systemic autoimmune diseases underly cardiac complications and may exacerbate development of heart disease. While this is well established for a range of autoimmune conditions, the clinical relevance of pressure-induced HF followed by autoimmunity is questionable. If the authors aim to address the clinical problem they refer to, autoimmunity needs to be induced before heart failure not after.

Answer: In hindsight we agree on this comment, and can assure the reviewer that this will be addressed in the future design of new experimental studies. The reason for the current experimental design was that we wanted to avoid the putative effects of arthritis-induced systemic inflammation on immediate peri- and postsurgical (cardiac) recovery. Also, as we believe an important question as to the vulnerability for RA patients to develop cardiac dysfunction/HF may be an already existing (subclinical) cardiac dysfunction, the current design would be of value.

+ The model used here is delayed-type hypersensitivity (which is not autoimmunity) with auto-Abs targeting an individual antigen (collagen). While no model of systemic autoimmunity is perfect, there are well accepted RA models that reflect human pathology better. Some of them can be time consuming and rather tedious to induce, so it is much appreciated that alternatives are being sought, but this needs to be reflected in the description of the study, starting with a precise title.

Answer: We agree that there are other models (like the CIA model) that may mimic its human counterpart better, especially in respect to autoimmunity and generalized arthritis (even though the specific autoimmunity in the human condition is quite different). Therefore, the putative value of the DTHA model we believe is that it may reflect the indirect consequence of arthritis-induced systemic inflammation. The differences in the various arthritis models are of importance and have now been addressed throughout the revised manuscript, starting with the title.

Minor comments:

+ It’s unclear why well established patient data should be repeated in a suboptimal sample size. Why not reference previous reports?

Answer: Our main objective in the analysis of the human RA serum was to relatively put into context the inflammatory responses seen in the DTHA model. As the ARCTIC study initially did not include age- and sex matched healthy controls, we needed to reanalyse a representative, though smaller subset of the ARCTIC study now including such controls.

Clarifications on this matter have been added to the text: Page 14, line 273-277 (in revised manuscript with track changes).

+ There are a few instances where ‘outliers have been excluded’. It is unclear what criteria were used to define outliers and how exclusion has been justified.

Answer: Outliers were excluded according to the ROUT method with a Q value of 1% (Q is the maximum desired false discovery rate). Information of this is added to the method section 2.8. In addition, we excluded TNF-values from two control mice even though they did not satisfy our exclusion criterion. In these mice both IL-6 and SAP values were defined as outliers which led us to consider these mice, or sampling of these mice, as erroneous. We also excluded an IL-6 value from the human control group due to similar considerations (both CRP and TNF values being outliers). This was not specified in the original manuscript, and all the individual excluded values are now specified in the relevant figure texts (Fig 2 and 6, and S2 Fig). The exclusion of these murine TNF and human IL-6 values did not influence the statistical results in respect to significance. The complete data including absolute values of outliers can be viewed in the raw data file attached (S1_File_Dataset).

+ There appears to be a small beneficial effect or the DTH-RA model on mild HF in some of the readouts? This might warrant discussion and further investigation considering that ‘fresh’ inflammation might in fact be beneficial if introduced at the right time.

Answer: We agree that some of the data comparing the impact of DTHA on cardiac function (i.e., SRe’) and weights (this including MRI estimated LV mass) leaves the impression of inflammatory beneficial effects on mild HF. However, in regards to function (Fig. 5B), this appears before inflammation is instigated. In respect to weights, an unexplainable difference in body weight (again appearing before induction of arthritis) may explain the differences. We now realize that the previous manuscript might have been unclear on this matter, and the current manuscript includes corrections in page 23, line 422-426 (in revised manuscript with track changes).

Reviewer #2: The authors have sought to investigate the link between RA and heart failure by assessing the impact of RA on cardiac function using a murine model of thoracic aortic constriction. As they indicate, RA patients are at increased risk of HF and the mechanisms driving this are still to be defined. This is therefore an important study and the manuscript is well written.

Comments:

1. It is interesting that the authors have chosen to study the effect of arthritis on the development of HF rather than the effect of arthritis on the development of cardiac dysfunction, although the rationale for this is explained. They have chosen an aortic constriction model driven by increased LV afterload. In the introduction both HFpEF and HFrEF is discussed and the literature suggests that RA patients may be particularly susceptible to HFpEF, which is driven by systemic inflammation (PMID: 34523821). However, the model chosen by the authors results in impaired EF. The authors are correct that there are few animal models to investigate HF in the context of RA and indeed models of diastolic dysfunction or HFpEF although there a few more examples in the literature that the authors may wish to discuss as the Pironti study is not the only experimental study that has investigated the cardiac consequence of RA as claimed in the discussion (PMID: 32208438; PMID: 33427610; PMID: 34526398). The authors might want to add a few lines to the discussion to distinguish between the different models of cardiac dysfunction and their relevance to the RA community.

Answer: We completely agree with the reviewer. These comments were also voiced by reviewer #1, and we refer to the answers given above. In short, these important studies have now been included and thoroughly discussed up against our data throughout the revised manuscript.

2. A timeline for the experimental protocol would be useful indicating when ORAB was initiated and then when RA was induced.

Answer: We agree that this would make the protocol clearer, and have uploaded a figure with a schematic illustration of the timeline in the supplemental material, see Figure S1.

3. Figure 1 panel A is this data mean +/- SEM (n=25) as indicated for panel B? Please add to panel A legend. No need to indicate ORAB in this legend as this data is not shown here.

Answer: This is both corrected in the manuscript.

4. Figure 2 (panel A-C) I am not sure what the authors mean by ‘due to space limiting reasons in the analysis……, 10 patients were randomised for analysis in each disease group’, yet there are 11 data points per group and it states that one outlier has been removed from each group. It is also not clear to me looking at the absolute values in supplemental figure 1, which patient(s) have been excluded as outliers as well.

Answer: From the databases we randomly selected 12 patients and controls. The demographic and some analytical data on these patients are all displayed in the supplemental table 1. Analysis of CRP was done separately, and only one of the datapoints were excluded as it met our exclusion criterion (in the control group). For the analysis of IL-6 and TNF we did have limitations in respect to the number of patients we could include in the 96-well plate. Thus, we chose to have 12 samples in the control group, and 10 samples in the disease groups (mild, moderate, severe). Of these, one outlier was removed in the control CRP group and one in the control TNF group. These outliers were from the same individual, and thus we chose also to exclude the corresponding IL-6 value. We also excluded one outlier in the moderate IL-6 group. We now realize that there is one discrepancy between the graphs displaying relative and absolute data (the excluded IL-6 datapoint in the control group). This has now been corrected (new absolute graph, S2 Fig). All the data, including the excluded data, can be seen in the raw data file attached (S1_File_Dataset).

The above considerations are added in detail in the now revised manuscript. See changes done in figure legend describing figure 2. Similar considerations/detailing have also been added for the murine data in the revised manuscript.

Also, why show the data as relative to control and then have another figure with the absolute values, rather than just showing absolute values in the main manuscript. There is some variance, as the authors have stated, but this is to be expected with human samples.

Answer: We appreciate the comments from reviewer #2. The purpose of this experiment was to compare inflammatory responses of mice subjected to experimental arthritis to human RA patients. We believe such comparisons between two different species are better visualized and represented using relative measures, as baseline values often differ (as they do here). We do, however, appreciate the importance of showing non-corrected data, and have thus included absolute presentations in the online supplement.

5. Were cytokine levels assessed in mice after the re-establishment of RA at days 15/18?

Answer: This is a valid question as it would add additional value to the interpretation of the DTHA model. We did consider to extend this part of our experiment, but chose not to. This decision was due to a (possibly erroneous) perception of minor additional informative value, but which would come with substantial additional financial costs and with an increase of mice needed.

6. Panel 2 (D-F) as above, I am not sure why 2 data points have been excluded for SAP and IL-6 and not TNF and again why data is shown as relative to control as well as absolute data in the supplemental file. Is this because statistical significance was not reached in the murine data when using absolute values? It states in the text, lines 303-306 that ‘Although non-significant due to variance in the data, we believe the data overall supports the notion that our experimental DTHA model…’ This leads me to believe that the absolute data shown in supplemental figure 2 is being referred to here, not the relative data shown in figure 2 where significance is shown. I agree with the author’s interpretation of the data and would just show absolute values as shown in the supplemental figure. I do not think you can correlate the relative data in mice to levels in patients, or that it is necessary to do so.

Answer: Regarding excluded datapoints, the rationale for these decisions has been thoroughly explained above (see also comments to reviewer #1). The relative and absolute analysis included the same values/datasets (with one error now corrected, see above) and the results are the same. However, we understand reviewer #2’s confusion as we have not displayed symbols of significance in the graph. This, and corresponding figure legend, has now been corrected.

Regarding the comparisons between the relative datasets of the inflammatory markers in mice and human, we now realize that our figure is confusing. The statistical annotation in Fig 2, D-F is not that of mouse vs human comparison. In order to visually appreciate the relative inflammatory responses on our arthritis model, we have inserted the average human values representing mild, moderate and severe disease. This we still believe is the most convenient way to display these data as the sole purpose of reanalyzing the clinical material, now including healthy volunteers, was to put the murine inflammatory responses into clinical context. We hope the altered text describing this (page 21, line 372-381 in revised manuscript with track changes) is more explanatory.

7. Line 303: It is not clear here that the text is now describing the murine data. Line 304 states CRP, rather than SAP.

Answer: We are describing the human data in line 303, which has now been specified and are hopefully now clearer (page 21, line 372 in revised manuscript with track changes).

8. Lines 407-411 in the discussion are not clear to me. I am not sure what the authors mean by time-amount mice with HF were exposed to ‘clinical relevant inflammation’.

Answer: We realize that this part of the discussion section is somewhat confusing and we have now tried to clarify it (see page 27, line 493-497 in revised manuscript with track changes).

Minor comment

Yellow lines are difficult to see clearly in the figures.

Answer: This has been changed to blue colour in fig 2, and pink colour in fig 3A, 4 and 5.

Legend for figure 3 panel B-D ends ‘mBSA, methylated’. Sentence is incomplete, but mBSA is not shown in the figure so this can be deleted – also in the legends for figures 4 and 5.

Answer: The reason for the incomplete sentence was a formatting error in the figure legend textbox. We chose to gather all the abbreviations at the end of the figure legend, and not at the end of each separate panel. Therefore, mBSA was explained after the figure legend of figure 3, panel B-D.

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

6 Jan 2022

Impact of delayed type hypersensitivity arthritis on development of heart failure by aortic constriction in mice

PONE-D-21-27012R1

Dear Dr. Tønnessen,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Michael Bader

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: (No Response)

Reviewer #2: The manuscript is much clearer. The discussion is particular is much improved and puts the research in context.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

10 Jan 2022

PONE-D-21-27012R1

Impact of delayed type hypersensitivity arthritis on development of heart failure by aortic constriction in mice

Dear Dr. Tønnessen:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Prof. Michael Bader

Academic Editor

PLOS ONE