We thank Tsung-Yuan Yang and colleagues for their interest in our findings on survival after COVID-19 associated organ-failure among SLE population. They raised two interesting questions on the method that we used.
First, they suspect a selection bias because we selected, for the unmatched analysis, patients still alive at D30 to measure the survival in the D30-D90 period while SLE patients had a lower mortality during the D0-D30 period. They stated that selecting patients based on what the next observation allocation is likely to be can lead to biased estimates. We agree with them, and, as we already wrote in the discussion section, “Such observation may be biased because patients with SLE are younger and more frequently female” which could explain the better prognosis during the D0-D30 period. Besides, we used D30 as a landmark not by choice, but because, in the matched analysis, (Figure 2) the Kaplan-Meier curves crossed at D30, and the proportional hazard assumption was therefore not respected. We did not drive conclusions from this unmatched analysis which was here mainly to show the importance of our matching procedure.
Second, they raised the concern that “the baseline characteristics between the two groups were not defined in this study”. We partly disagree on this comment. As a matter of fact, we presented in Table 1 (unmatched analysis) and in Table 2 (matched analysis) the baseline characteristic of our studied populations. We presented all the data that we had regarding the simplified acute physiology score 2 (SAPS 2) which is also a classification tool for individual’s risk mortality in hospital (1) and an analogue of APACHE 2 used in France. However, we acknowledge that this score is only collected in intensive care units (ICUs), so it was available only for patients admitted in these units. As presented in table 2, and even though SAPS 2 was not part of the matching variables, this score was distributed similarly between SLE and non-SLE matched populations.
Overall, we provide evidence that, after considering age, sex, and several comorbidities, SLE patients have a late-onset poor prognosis after COVID-19 AOF.

References:
1-Le Gall J, Lemeshow S, Saulnier F. A New Simplified Acute Physiology Score (SAPS II) Based on a European/North American Multicenter Study. JAMA. 1993;270(24):2957–2963.

We read with interest the Viewpoint article by Braun and Landewé regarding post-hoc analysis of back pain in trials of IL-23 inhibitor therapy in patients with peripheral psoriatic arthritis (PsA) [1]. Indeed, we share their concerns regarding study design, the use of outcome measures developed for axial spondyloarthritis (axSpA) and, most importantly, the attribution of the diagnostic label “physician-reported spondylitis” in these patients. In addition to the issues eloquently outlined in the article, it is important to be aware that the pre-test probability of inflammatory disease being directly responsible for back pain is likely much lower in patients with PsA who are older, and therefore more likely to have mechanical or non-specific back pain, than people presenting with axSpA. In other words, even before doing any test, a middle-aged person with PsA, as represented in most phase III PsA clinical trials, is more likely to have non-inflammatory than inflammatory back pain. These “causes” of back pain do of course co-exist and are not easily distinguished by clinical or imaging assessments. For example, disc and degenerative spinal disease can lead to apparent inflammatory features, such as bone marrow oedema on magnetic resonance imaging [2, 3] that are likely a secondary response to altered biomechanical stresses rather than primary inflammatory disease and therefore unlikely to be responsive to biologic therapies. Furthermore, imaging data on the prevalence and natural history of spinal involvement in PsA are limited. We therefore agree that more research is required to both define inflammatory axial involvement in PsA (for which global initiatives are ongoing, as outlined by Braun and Landewé) and to evaluate the effect of specific therapies on this component of psoriatic disease. The latter will need to be tested in appropriately recruited PsA populations and will likely require comparison between active therapies with similar efficacy in both peripheral joint and skin disease to mitigate for the bystander effect resulting from improvements in these non-axial domains.

However, until this is resolved, clinicians are faced with a dilemma when choosing a biologic therapy for PsA patients with significant spinal symptoms. Indeed, where spinal symptoms are the sole or dominant musculoskeletal complaint in a person with psoriasis, there is clearly a requirement to determine, as far as possible, whether or not there is likely to be active inflammatory axial disease to warrant biologic therapy. This should be done using a combination of clinical, imaging and laboratory assessment, while also taking into account context and likelihood. However, in PsA patients with spinal symptoms who also have significant active peripheral joint disease that warrants biologic therapy in its own right, should IL-23 inhibitors be avoided on the basis of the negative data from for these agents in axSpA trials [4, 5]? For now, we believe that in light of the difficulty in determining the nature of the axial involvement in PsA, treatment decisions should be based on considerations relating to peripheral musculoskeletal disease, extent of cutaneous psoriasis, previous therapies, extra-articular manifestations, comorbidities and safety to select the most appropriate therapy for these patients. While we agree with Braun and Landewé that the current post-hoc analyses do not prove that IL-23 inhibitors are efficacious for inflammatory axial disease in PsA, we suggest that, until further robust evidence is available, the presence of axial symptoms, whether inflammatory or non-inflammatory, should not be considered a reason to avoid IL-23 inhibitor therapy, if that is perceived to be otherwise the most appropriate and safest choice for that individual’s peripheral musculoskeletal and cutaneous disease.

REFERENCES
1. Braun J, Landewé R. No efficacy of anti-IL-23 therapy for axial spondyloarthritis in randomised controlled trials but in post-hoc analyses of psoriatic arthritis-related ‘physician-reported spondylitis’? Ann Rheum Dis Published Online First: 16 October 2021. doi: 10.1136/annrheumdis-2021-221422
2. Tsoi C, Griffith JF, Lee RKL, et al Imaging of sacroiliitis: Current status, limitations and pitfalls. Quant Imaging Med Surg 2019;9:318-35
3. Berthelot JM, le Goff B, Maugars Y, et al. Sacroiliac joint edema by MRI: Far more often mechanical than inflammatory? Joint Bone Spine 2016;83:3-5
4. Deodhar A, Gensler LS, Sieper J, et al. Three multicenter, randomized, double-blind, placebo-controlled studies evaluating the efficacy and safety of ustekinumab in axial spondyloarthritis. Arthritis Rheumatol 2019;71:258–70.
5. Baeten D, Østergaard M, Wei JC-C, et al. Risankizumab, an IL-23 inhibitor, for ankylosing spondylitis: results of a randomised, double-blind, placebo-controlled, proof-of-concept, dose-finding phase 2 study. Ann Rheum Dis 2018;77:1295–302.

ACKNOWLEDGEMENTS
HM-O is supported by the National Institute for Health Research (NIHR) Leeds Biomedical Research Centre. The views expressed are those of the authors and not necessarily those of the NIHR or the Department of Health and Social Care.

We read with great interest the recent study by Kai-di Wang and colleagues1, published in the Annals of Rheumatic Disease, which showed LRP4 was isolated as a novel target of digoxin, and deletion of LRP4 abolished digoxin’s regulations of chondrocytes. We appreciate this meaningful research very much, and we believe that this study has significant guiding for providing new insights into the understanding of digoxin’s chondroprotective action and underlying mechanisms, but also present new evidence for repurposing digoxin for osteoarthritis (OA). However, we have some questions to discuss with the authors except for the limitations the authors mentioned in the study.
As all we know, the authors used two cardenolides to conduct the experiments, ouabain and digoxin. We can know that ouabain and digoxin can both enhance chondrogenesis and stimulate chondrocyte anabolism from the result of Figure 1. More importantly, we found that the relative staining level in ouabain treated groups was much higher compared with that in digoxin group (Figure 1B). Moreover, the mRNA expressions of transcriptional levels of chondrogenic marker genes, such as COL2A1, Comp, ACAN, SOX5, SOX6, and SOX92-4 were also much higher compared with that in digoxin group (Figure 1C). The similar situation was showed in Figure 1D-G, which may indicate ouabain better enhance chondrogenesis and stimulate chondrocyte anabolism than that of digoxin. In addition, we may also speculate ouabain better protect against OA in a surgically induced model in vivo than that of digoxin, especially from Figure 2H and Figure 2I. However, the study emphasized the importance of digoxin instead of ouabain from the beginning to the end. Therefor we are wondering what the role of ouabain played in the whole experiment procedure compared to digoxin on earth.
On the other hand, Figure 4 demonstrated that digoxin use was associated with reduced risk of knee or hip OA-associated joint replacement among patients with atrial fibrillation. This is easy to understand. However, we want to know how many OA patients with atrial fibrillation in the world. What’s the percentage this kind of patients in the whole OA patients?
This is why we are more interest in digoxin used in OA patients without atrial fibrillation. Whether digoxin can have the same protection among OA patients without atrial fibrillation.What’s more, What’s the side effect when using digoxin to OA patients with normal heart function?
We would like to point out that since ouabain and digoxin can both enhance chondrogenesis and stimulate chondrocyte anabolism to protect against OA. We suppose whether we can use digoxin combined with ouabain at the same time to treat with OA patients. The most reason why this hypothesis we have is that we can perhaps reduce the side effect by using lower dose of digoxin, but do not affect the function of protecting against OA patients. And digoxin plus ouabain with different doses group can be planned to further experiments. Maybe we can find a better treatment to OA patients.
References
1 Wang KD, Ding X, Jiang N, et al. Digoxin targets low density lipoprotein receptor-related protein 4 and protects against osteoarthritis. Ann Rheum Dis. 2021 De1: annrheumdis-2021-221380.
2. Nenna R, Turchetti A, Mastrogiorgio G, et al.COL2A1 Gene Mutations: Mechanisms of Spondyloepiphyseal Dysplasia Congenita. Appl Clin Genet. 2019;12:235-238.
3. Lin EA, Liu C-J. The role of ADAMTSs in arthritis. Protein Cell. 2010;1:33–47.
4. Liu C-ju, Zhang Y, Xu K, et al. Transcriptional activation of cartilage oligomeric matrix
protein by SOX9, SOX5, and SOX6 transcription factors and CBP/p300 coactivators.
Front Biosci.2007;12:3899–910.

We read with great interest the article by Corbera-Bellalta et al. entitled “Blocking GM-CSF Receptor α with mavrilimumab reduces infiltrating cells, pro-inflammatory markers, and neoangiogenesis in ex-vivo cultured arteries from patients with giant cell arteritis” (1). We believe that the study is of great importance because it demonstrates that blocking the GM-CSF pathway alleviates crucial pathological hallmarks of giant cell arteritis (GCA). These hallmarks include leukocyte infiltration, production of pro-inflammatory cytokines, tissue destructive matrix metalloproteinases (MMP’s), and neoangiogenesis. Additionally, the recently reported preliminary data of the first phase II clinical trial of mavrilimumab in combination with a 26-week glucocorticoid (GC) taper in GCA patients is very promising (2). The primary end point, being the difference in the time to first relapse between mavrilimumab treatment and placebo was achieved (p=0.0263). Moreover, the sustained remission rate at 26 weeks was higher in the mavrilimumab vs placebo group (83.2% vs 49.9%, respectively, p =0.0038).

Recently, we reported on a distinct CD206+ macrophage subset that produces YKL-40 and MMP-9 in GCA affected vessels (3, 4). We proposed a pathogenic model in which these CD206+ macrophages play major roles in fueling leukocyte infiltration, vascular destruction, and neoangiogenesis. Furthermore, we showed that these CD206+/MMP-9+/YKL-40+ tissue destructive and proinflammatory macrophages are induced upon GM-CSF signaling. In agreement with our findings, Corbera-Bellata et al. show that blockade of the GM-CSF Receptor α with mavrilimumab resulted in marked reduction in expression of CD206 (transcript) and MMP-9 (transcript & protein levels), leukocyte infiltration and neoangiogenesis.

We would like to put forward that YKL-40, also known as human cartilage-glycoprotein 39 or Chitinase 3-like 1/CHI3L1), could be a relevant biomarker to aid GCA patient stratification for treatment with mavrilimumab. YKL-40 is highly expressed by macrophages in inflammatory conditions and its overexpression is highly driven by GM-CSF (4, 5). Additionally, YKL-40 has been reported to be one of the upstream signals for MMP-9 expression by macrophages involved in tissue destruction and in the promotion of neovessel formation, two crucial processes in the pathogenesis of GCA (6, 7). Our group has previously reported that serum levels of YKL-40 are elevated at diagnosis and remain elevated during GC treatment in GCA patients (4, 8). Furthermore, higher levels of YKL-40 at diagnosis predicted a longer duration of GC treatment (8). In line with elevated YKL-40 levels, myeloid cells are also elevated in the circulation of GCA patients at diagnosis and during GC treatment (9). In tissue, macrophages remain present in the vessel wall after GC treatment (10). Thus, it appears that macrophages producing YKL-40, likely skewed by GM-CSF, are insufficiently targeted by GC treatment. As mavrilimumab treatment blocks the GM-CSF pathway and thereby affects the myeloid compartment, patients with high YKL-40 levels at baseline may especially benefit from mavrilimumab treatment. Future studies should evaluate whether serum levels of YKL-40 predict the response to mavrilimumab. Furthermore, it would be interesting to determine to what extent mavrilimumab is able to suppress the persistent YKL-40 response in patients with GCA. Taken together, we propose that serum YKL-40 measurements could be of interest for precision medicine and monitoring treatment efficacy of patients with GCA in the context of mavrilimumab treatment.

References:
1. M. Corbera-Bellalta, et al., Blocking GM-CSF receptor α with mavrilimumab reduces infiltrating cells, pro-inflammatory markers and neoangiogenesis in ex vivo cultured arteries from patients with giant cell arteritis. Ann. Rheum. Dis., annrheumdis-2021-220873 (2022).
2. M. C. Cid, et al., Mavrilimumab (anti GM-CSF Receptor α Monoclonal Antibody) Reduces Time to Flare and Increases Sustained Remission in a Phase 2 Trial of Patients with Giant Cell Arteritis [abstract]. Arthritis Rheumatol. 2020; 72 (suppl 10).
3. W. F. Jiemy, et al., Distinct macrophage phenotypes skewed by local granulocyte macrophage colony‐stimulating factor (GM‐CSF) and macrophage colony‐stimulating factor (M‐CSF) are associated with tissue destruction and intimal hyperplasia in giant cell arteritis. Clin. Transl. Immunol. 9 (2020).
4. Y. van Sleen, et al., A Distinct Macrophage Subset Mediating Tissue Destruction and Neovascularization in Giant Cell Arteritis: Implication of the YKL-40 - IL-13 Receptor α2 Axis. Arthritis Rheumatol. (2021) https:/doi.org/10.1002/ART.41887 (July 28, 2021).
5. J. S. Johansen, K. S. Krabbe, K. Moller, B. K. Pedersen, Circulating YKL-40 levels during human endotoxaemia. Clin. Exp. Immunol. 140, 343–348 (2005).
6. S. Libreros, R. Garcia-Areas, P. Keating, R. Carrio, V. L. Iragavarapu-Charyulu, Exploring the role of CHI3L1 in “pre-metastatic” lungs of mammary tumor-bearing mice. Front. Physiol. 4, 392 (2013).
7. R. Watanabe, et al., MMP (matrix metalloprotease)-9-producing monocytes enable T cells to invade the vessel wall and cause vasculitis. Circ. Res. 123, 700–715 (2018).
8. Y. van Sleen, et al., Markers of angiogenesis and macrophage products for predicting disease course and monitoring vascular inflammation in giant cell arteritis. Rheumatology (Oxford). 58, 1383–1392 (2019).
9. Y. van Sleen, et al., Leukocyte Dynamics Reveal a Persistent Myeloid Dominance in Giant Cell Arteritis and Polymyalgia Rheumatica. Front. Immunol. 10 (2019).
10. J. J. Maleszewski, et al., Clinical and pathological evolution of giant cell arteritis: A prospective study of follow-up temporal artery biopsies in 40 treated patients. Mod. Pathol. 30, 788–796 (2017).