The role of PCSK9 in inflammation, immunity, and autoimmune diseases

ABSTRACT

Introduction

Statins have pleiotropic effects, being both anti-inflammatory and immunomodulatory. Proprotein convertase subtilisin kexin 9 (PCSK9) targets the low-density lipoprotein receptor (LDLR), which increases LDL levels due to the lower expression of LDLR.

Areas covered

Inhibition of PCSK9 by the use of antibodies represents a novel principle to lower LDL levels. LDL may have other properties than being a cholesterol carrier but is well established as a risk factor for cardiovascular disease and atherosclerosis. In atherosclerosis, the plaques are characterized by activated T cells and dendritic cells (DCs), dead cells, and OxLDL. The latter may be an important cause of the inflammation typical of atherosclerosis, by promoting a proinflammatory immune activation. This is inhibited by PCSK9 inhibition, and an anti-inflammatory type of immune activation is induced. OxLDL is raised in systemic lupus erythematosus (SLE), where both CVD and atherosclerosis are much increased compared to the general population. PCSK9 is reported to be associated with disease activity and complications in SLE. Also in other rheumatoid arthritis, PCSK9 may play a role.

Expert opinion

PCSK9 has pleiotropic effects, being implicated in inflammation and immunity. Inhibition of PCSK9 is therefore interesting to study further as a potential therapy against inflammation and autoimmunity.

KEYWORDS:

• PCSK9
• LDL
• SLE
• immunology
• atherosclerosis

1. Background

Proprotein convertase subtilisin/kexin type 9 (PCSK9) is a serine protease. It is a ligand for the low-density lipoprotein receptor (LDLR) and a major player in cholesterol homeostasis, with LDL being an important carrier of cholesterol [1].LDL binds to LDLRs and is internalized, which clears LDL from the circulation. After digestion in the lysosome, LDLR is recycled to the cell surface, where it continues its LDL clearance. When PCSK9 binds to LDLR, lysosomal degradation inside the cells is induced, which leads to decreased LDLR expression on the cell membrane. PCSK9 also inhibits the recirculation of LDLR on cells as hepatocytes through inducing the degradation of LDLR in the lysosomes. Together, this leads to less expression of LDLR and increased LDL levels [2].

The role of PCSK9 in the regulation of the LDLreceptor was demonstrated in genetic studies, where variants were identified, which are related to low PCSK9 activity and also decreased risk of cardiovascular disease (CVD) [3,4]. Also genetic variants which lead to high PCSK9 activity have been identified, which causes high LDL-levels and increased CVD risk [5].

The identification of PCSK9 and its role in LDLR and LDL regulation led to the development of novel drugs targeting LDLR by inhibition of PCSK9 to lower LDL levels. Currently, two monoclonal antibodies are used therapeutically to lower LDL, and in clinical trials, LDL levels were strongly reduced as well as the risk of CVD [6–10].

2. Low-density lipoprotein

Atherosclerosis and its consequence cardiovascular disease CVD) represent major causes of morbidity and mortality and their major risk factors have been known for a long time. Among modifiable ones, hypertension, smoking, diabetes, and high low density lipoprotein (LDL) have been known for a long time. In addition, many others have been discussed, including inflammation, though to qualify as a risk factor, treatment of it should reduce the risk, and as yet, there is no specific treatment against atherosclerosis and CVD targeting inflammation. Nonmodifiable risk factors include male sex and age. Atherosclerosis has been reported in both ancient Egyptian mummies [11,12] and in the famous Ice-man, Ötzi, who lived his life mainly as a hunter and gatherer some 7000 years ago, indicating that atherosclerosis is not only related to modern life including a sedentary life style [13].

Atherosclerosis is an inflammatory disease process, characterized by the accumulation of dead cells, LDL (which has undergone enzymatic modification and/or oxidation), and activated immune competent cells in the the artery wall. The cells include T-cells and dendritic cells (DCs), macrophages, and others [14,15].

While LDL is well established as a risk factor, the underlying mechanisms are not fully elucidated. Of note, the LDL accumulating in the atherosclerotic plaques of arteries is modified, mainly oxidized. Currently, a major hypothesis about why LDL is atherogenic is that this is caused by the oxidation of LDL [16]. This can occur through different mechanisms, including by enzymes such as phospholipase A2, which is expressed in plaques, and by chemical modification, e.g. by metal ions [15,17].

OxLDL is a complex compound, and several different derivatives generated during LDL oxidation can cause damage and increase atherosclerosis. Oxidized phospholipids (OxPL) where phosphorylcholine (PC) is an important epitope are of major interest, and so is another phospholipid, lysophosphatidylcholine (LysoPC), and also malondialdehyde (MDA). Also the carrier protein in LDL, apoB, has been implicated, and apoB can be modified by oxidation or enzymatically. These possibilities are nonmutually exclusive, though much evidence has been accumulated supporting an important proinflammatory role of OxPL and the other lipid-related compounds. OxLDL is proinflammatory, inducing immune activation of both monocyte and T cells [18–20] and we reported in the mid-90s that PC is a major cause of OxLDL-induced immune activation [21]. These studies were followed up recently, with findings indicating that OxLDL activates dendritic cells, which promote proinflammatory T cell activation in humans, including in atherosclerotic plaques and among patients with systemic lupus erythematosus (SLE) where the risk of CVD is very high [22–25]. Also MDA carried by human serum albumin has similar properties to OxLDL in this context [26].

PC exposed on oxLDL and also on other PC carriers in the circulation is raised in high-risk individuals, e.g. with the presence of hypertension or with systemic lupus erythematosus (SLE) [27–29].

LDL has properties in addition to being a cholesterol carrier and being proinflammatory after inflammation. LDL may play a role in protection against infections and potential mechanisms including its role as scavenger, e.g. in relation to lipopolysaccharide. This notion is supported by animal experiments, where LDL protects against infection [30–32].

In patients with kidney failure, mortality is high because the risk of CVD and fatal infections is very high. Here LDL-lowering therapy is not known to be very effective, in spite of high LDL levels, and interestingly, LDL levels have been reported to be inversely associated with mortality due to infections, supporting the notion that LDL may at least in some circumstances act against infections [33]. In line with the notion that LDL may be protective in severe infections as in sepsis is a recent report where low LDL was strongly associated with mortality among sepsis patients in intensive care units, while prehospitalization LDL was not [34].

It is also interesting to note that patients with familial hypercholesterolemia in a study of a pedigree from Netherlands, did not have increased mortality in the 19^th century, which was the case later on, FH being known to increase CVD and mortality from CVD. This finding is also compatible with a protective role in circumstances where infections represent a high risk of morbidity and mortality [35]. The possibility that FH has been an advantage from an evolutionary point of view during times with high mortality from infections, including death among infants, deserves further attention. Still, the role of PCSK9 during human evolution remains obscure, though the different mutations leading to both high and low levels of the enzyme (and LDL) are of interest in relation to the different roles played by LDL.

In the context of PCSK9, it is interesting to note that statins, which decrease cholesterol synthesis by the inhibition of the enzyme HMG-CoA reductase, have been reported to have other potentially protective effects than LDL-lowering, including anti-inflammatory and also immune modulatory effects. [15,25] Of note, PCSK9 and LDLR gene expressions are both regulated by sterol regulatory element-binding protein 2 (SREBP2) and when intracellular levels are decreased as in statin therapy, SREBP2 activation promotes PSCK9, which may decrease statin therapy effects to a variable extent [36,37].

Paraoxonases, especially paraoxonase-1 (PON1), may play an important role in atherosclerosis and lipid peroxidation. PON1 binds to high-density lipoprotein (HDL) and breaks down lipid peroxides in LDL, which may be part of the atheroprotective properties of HDL. However, little is known about the potential interactions between PCSK9 and PON1 (or other paraoxonases) [38].

3. Potential pleiotropic effects of PCSK9

3.1. General findings

Different lines of evidence indicate that PCSK9 may have properties beyond its effects on LDLR and LDL levels. In a prospective cohort started in the early 90s from Åkersberga north of Stockholm where above 4000 60-year-olds were included, PCSK9 levels were associated with increased risk of future CVD independent of LDL levels [39]. High levels of PCSK9 have been reported to be associated with increased atherosclerosis progress independent of LDL levels [40].

Of note, there is a variability in results, and other large studies fail to demonstrate an association between PCSK9 levels at the baseline and future risk of CVD or the vascular structure [41,42]. There could be important differences between populations in relation to PCSK9 risk, e.g. dependent on age and gender.

Interestingly, treatment with PCSK9 inhibitors in patients with familial hypercholesterolemia decreases the levels of platelet-activating factor (PAF) and its precursors, e.g. PAF-like lipids [43]. This is interesting since PAF and PAF-like lipids in OxLDL cause a significant activation of immune competent cells to produce IFN-gamma [21]. PCSK9 inhibition could thus be beneficial also through this mechanism, not directly related to its lipid-lowering ability.

3.2. PCSK9 in atherosclerosis and inflammation

In contrast to the many reports about pleiotropic effects of statins, both related to inflammation and immunity, the knowledge about such properties of PCSK9 and PCSK9 inhibitors is relatively scarce. One earlier example from cell experiments is the observation that smooth muscle cells secrete PCSK9, which then decreases LDLR expression in macrophages [44]. Smooth muscle cells change the phenotype in atherosclerosis and are likely to play an important role in the buildup of plaques when they migrate into the intima from the media.

Interesting effects on scavenger receptors have also been reported. In general, OxLDL is taken up by macrophages in the atherosclerotic plaques through different scavenger receptors, including CD36, SRA, and LOX-1. These macrophages become inert instead of removing its load of OxLDL and develop into foam cells in the plaques where they stay and eventually die, and thus constituting a large proportion of the necrotic core, which is typical of atherosclerosis. Even though the role of scavenger uptake and atherogenesis is complex, most evidence indicate that this is atherogenic, e.g. promotes atherosclerosis, at important stages of plaque development. [15]

PCSK9 induces an increase in CD36, SRA, and LOX-1 at gene and protein levels in mouse macrophages, which leads to increased OxLDL uptake. In macrophages lacking these scavenger receptors, this increased uptake did not occur, suggesting an effect on foam cell generation by PCSK9 in atherosclerosis [45].

In line with this, proinflammatory responses in macrophages have been reported to be induced by PCSK9 [46].

There are also animal experiments supporting pleiotropic effects in relation to PCSK9. By the use of chimeric mice expressing PCSK9 in macrophages, a proinflammatory effect in lesions with increased infiltration of monocytes and macrophages was reported, which supports the notion that PCSK9 has pleiotropic effects, promoting inflammation in atherosclerosis. [47]

3.3. PCSK9 in immunity

In specific, T-cell-mediated immune reactions, DCs are generally believed to play a major role, by presenting antigens to T cells and also influencing the type of T cell reaction that ensues. A discussed, T cells are present in atherosclerotic plaques, which was reported in the 80s [48], and likewise, DCs are present and are likely to play an important role in different stages of atherosclerosis development, not least at late stages, when plaques may be damaged, and may rupture, which leads to CVD [49–51]. As expected in immune reactions of this type, DC and T cells co-localize in plaques, and interestingly, often in the vicinity of sites where plaques rupture [52].

As discussed, oxLDL is an antigen that may play a major role in immune activation in atherosclerosis, even though there are other, nonmutually exclusive possibilities. [14,15] One is heat shock protein 60, which is immunogenic and is induced by OxLDL. It is possible that OxLDL to some extent activates T cells through the induction of HSP60, which is presented as antigen by DCs [23,53].

Animalexperiments support that T cells may promote atherosclerosis, e.g. transfer of CD4+T Cells to micemodels of atherosclerosis promote the disease condition. Also abrogation of

of transforming growth factor-beta (TGF-ß) is atherogenic [54,55]. This notion is also supported by human data on T cells and CVD [56].

We recently demonstrated that statins have immunomodulatory properties and promote an anti-inflammatory T cell response to OxLDL-induced DC-T cell activation and also determined specific intracellular pathways [25]. This finding provided a basis for our studies on PCSK9 in similar ex vivo systems, where T cells and DCs from atherosclerotic plaques and patients with symptoms of threatening CVD are operated upon. We reported in immunological studies that oxLDL induced PCSK9 in DCs and also induced DC maturation. Silencing of PCSK9 inhibited OxLDL-induced T cell activation by DCs and also promoted an anti-inflammatory phenotype, where OxLDL-induced polarization to Th1 and Th17 cells was reversed. Also proinflammatory cytokines such as TNF-α, IL-1β, and IL-6 were inhibited by PCSK9 silencing (with mock silence as control). Instead, TGF-β and IL-10 and T regulatory cells were induced. Specific underlying mechanisms involved miRNA. OxLDL induced let-7 c, miR-27a, miR-27b, and miR-185, while silencing of PCSK inhibited miR-27a and let-7c. The effect of statins on mi-RNA differed somewhat from PCSK9-inhibition [25]. Also SOCS1 expression, induced by oxLDL, was enhanced by silencing of PCSK9. These findings involved ex vivo cell culture systems from both atherosclerotic plaques and healthy donors [22]. Similar results were obtained with DCs and T cells from SLE patients in another study [57]. Taken together, PCSK9 could thus have specific immunomodulatory effects on OxLDL-mediated immune activation in the plaque, through specific mechanisms, suggesting an immunological role in addition to LDL lowering, to ameliorate atherosclerosis.

In line with this, a reduced number of Th17 T cells was reported in mice lacking PCSK9. The authors suggested that PCSK9 has a role in T cell programming and also through this mechanism, in atherosclerosis development [58].

Even though beyond the scope of this review, a recent article in Nature gives support to a role of PCSK9 in T cell biology, here in cancer. In this elegant study on cancer immunology, PCSK9 was demonstrated to boost the response of tumors to immune checkpoints therapy, independent of lipid lowering. Deletion of PCSK9 in the mouse model prevents tumor growth through T cells. One mechanism is to regulate surface MHC I levels and promote intratumoral T cell infiltration. Also, antibodies against PCSK9 acted in synergy with check point immune therapy against cancer tumors [59].

3.4. PCSK9 in autoimmune and inflammatory diseases

SLE is often referred to as a prototypic autoimmune disease. The risk of CVD in SLE is very high, and it has therefore been suggested that SLE could give important information about the role of the immune system in atherosclerosis, in a similar way as chronic kidney disease (CKD) may be discussed [29,60–63]. In general, a combination of traditional and nontraditional risk factors accounts for the increased risk of CVD in SLE. This increased risk could be caused by increased thrombosis propensity, and many SLE patients have high levels of thrombogenic antiphospholipid antibodies (aPL) [29]. Hypertension is important among the traditional risk factors. This is the case also with dyslipidemia, though in SLE, the dyslipidemia is characterized by high triglycerides, low high-density lipoprotein (HDL) but LDL is sometimes decreased in active disease [64].

Among the nontraditional ones, we reported that low levels of anti-PC are independently associated with increased risk of CVD. Anti-PC, especially IgM and IgG1 anti-PC, is associated with protection against CVD and also atherosclerosis progress and has several potentially protective properties, including anti-inflammatory, inhibiting the obnoxious properties of OxLDL and its derivatives [15,62,63]. Not only CVD but also atherosclerosis is increased in SLE, but in most studies, this is atherosclerosis as determined by atherosclerotic plaques, not the general intima-media thickness (IMT) [15,29,62–64].

As mentioned, thrombogenic aPLs are common in SLE and can cause thrombosis both by a direct mechanism, targeting the endothelium, and indirectly, through competing out Annexin A5 from binding damaged endothelium. AnnexinA5 has anti-atherogenic and anti-inflammatory properties and binds not only phosphatidylserine (exposed on dead cells) but also OxLDL where it inhibits proinflammatory effects [23,24,65]. However, aPL are not associated with atherosclerosis per se [29].

Relatively little is known about PCSK9 in SLE. It was reported that PCSK9 levels are raised in SLE and also associated with both nephritis and measures of atherosclerosis (IMT) [66].

We reported that PCSK9 levels are associated with both SLEDAI and SLAM, two measures of disease activity in SLE, but only nonsignificantly raised among SLE patients as compared to controls. We did not determine any significant association with atherosclerosis, though, not with IMT and not with prevalence of plaques, including echolucent (potentially vulnerable) plaques. Still, levels were strongly associated with CVD, though not after adjustment for age [62]. Eventhough the findings point in the same direction, suggesting an association with inflammation and inflammation-related complications in SLE, there were differences. These may be explained by different age in the studies (mean age in our 48 vs 33 in the other) and disease activity (SLEDAI 2 vs 6), and PCSK9 levels were higher in the other study. Furthermore, controls were population based in our study. Further studies are thus needed to establish a potential role of PCSK9 and its inhibition in SLE.

PCSK9 could be one factor playing a role potentially independent of lipid levels in these conditions but promoting inflammation [67]. Another interesting question is if PCSK9 interacts with treatment by biologics or small molecule-related medication in autoimmune disease, though to date, little is reported.

Wealso determined that OxLDL-induced DC and T cell activation is stronger among SLE patients than controls, and that silencing of PCSK9 promoted an anti-inflammatory phenotype with an increase in T regulatory cells and inhibition of DC activation and maturation. It is interesting to note, however, that the inhibition by PCSK9 silencing was not identical to the effects by statins, since there were differences in miRNA usage among others [22,25].

We also determined that DCs from controls expressed less PCSK9 than those from SLE patients. In the local environment of plaques (and other tissues), overexpression of PCSK9 could thus increase local inflammation.

In line with this, another study reported that SLICC, which is a damage index, and SLEDAI, which is an index of disease activity in SLE, were both independently associated with PCSK9-levels [68].

These findings imply that PCSK9 could be of importance in SLE by different though related mechanisms though the details of these are not known. OxLDL is raisedin SLE, induces PCSK9 in DCs from SLE patients, and activates DCs and T cells more in SLE than among controls. Further, OxLDL-induced immune activation is ameliorated by PCSK9 inhibition and T regs (which are believed to be protective in SLE) are promoted. An interesting possibility is therefore that PCSK9 inhibition could be beneficial for SLE patients, especially those with high disease activity, but also to inhibit complications like nephritis and increased atherosclerosis.

Antiphospholipid antibodies (aPLQ) are commonly found in SLE, which also can represent a primary disease. Manifestations include CVD, but also miscarriage. aPL may be one important risk factor for CVD in SLE, but does not appear to be related to increased atherosclerosis.

Ina recent development, PCSK9 has been implicated in aPL. First, in a genetic study, the LDLR gene was associated with the development of thrombosis in aPLA. Furthermore, there was also a significant association with the PCSK9 gene in this patient group [69]. In another study, it was determined that PCSK9 binds to APOH, which is reported to involved as an antigen in aPLA-patients [70].

The prevalence of rheumatoid arthritis (RA) is 0.5–1% and in addition to being a burden on individuals, this implies that it is also costly for society[71]. Due to the introduction of biologicals, the prognosis has improved significantly, and tumor necrosis factor α (TNF-α) antagonists were the first ones introduced, typically in combination with methotrexate and other established disease modifying medications [72]. In the context of PCSK9-inhibition therapy, it is interesting to note that combination therapy is very common in RA. There are other biologics with different types of cytokine-inhibitory and anti-inflammatory effects. Still, around 30% of patients do not respond to treatment [73,74].

As in several other autoimmune conditions, patients with RA are at increased risk of atherosclerosis and its complications, which may be inhibited by biologics [75–77].

We reported that PCSK at the baseline among TNF-α antagonist-treated RA patients is negatively associated with disease activity, as determined by standard methods (DAS28) after 3 months, 6 months, and 12 months. RA patients who had PCSK9 levels in the lowest quartile had 4 times increased chance of being in remission, with no signs of active disease. An interesting possibility is therefore that PCSK9 could be used as a risk marker for being nonresponder to biologics in RA [78].

In RA, the synovium in the joints is infiltrated by activated macrophages, neutrophils, and lymphocytes and are directly involved in disease development. Also the synovial cells show signs of activation and participate in the ongoing inflammation [79,80].

We also demonstrated that PCSK9 at physiological concentrations induces TNF-α and IL-1-ß from macrophages, e.g. two major players in the inflammation in RA, proven by the therapeutic effects of biologics inhibiting these agents. Further to this, PCSK9 induced MCP-1 from human synoviocytes in similar cell culture systems.

MCP1 is believed to be involved in RA pathogenesis, one mechanism being the recruitment of macrophages [81] and inhibition of MCP-1 ameliorate arthritis in a rat models [82]. In line with this, levels of MCP-1 is raised in RA-patients [81]. One possibility is therefore that PCSK9-induced MCP-1 could contribute to the recruitment of mononuclear leukocytes [78].

Induction of TNF-α and IL-1-ß from macrophages and MCP-1 from synoviocytes was inhibited by antibodies against PCSK9. The possibility that PCSK9 inhibition could be of therapeutic value in RA (also in relation to CVD, not only RA per se) thus deserves further consideration, maybe especially in patients with high levels of PCSK9 [78].

Of note, foam cells with OxLDL have been detected in RA, suggesting common underlying mechanisms in RA and atherosclerosis [31,83]. OxLDL is raised in RA and is associated with CVD in RA [32,84]. Induction of Tregs is generally believed to be beneficial in RA [85]. If OxLDL plays a role also in RA, the previously discussed mechanisms where PCSK9 ameliorates proinflammatory and immune activating properties of OxLDL could thus also play a role in this disease condition.

In psoriasis, evidence is emerging that PCSK9 may play a role. In one study, it was demonstrated that three months of monotherapy with methotrexate reduced PCSK9 levels and PCSK9 may be a marker of psoriasis [86].

4. Expert opinion

Lowering of LDL levels to decrease the risk of CVD and atherosclerosis through medication is widely used and plays a role in the age-matched decrease in CVD that has occurred during the last decades. Statins were a clear step forward, and now several new versions of this drug category are used. PCSK9 inhibition represents another principle to lower LDL and is now mostly complementary, since statins remain the first in line treatment. The history of PCSK9 and its inhibition is very interesting also from a more general point of view, with the combination of elegant genetic studies, with experiments. The presence of different mutations, leading to both high and low LDL levels is striking, and its role from an evolutionary point of view should be studied further. In my opinion, it indicates that LDL is far more than a transporter of cholesterol, for example, it may be of importance in the body´s countermeasures against infections, which historically have been a huge problem and major cause of morbidity and mortality and still is in parts of the world. Genotypes causing high LDL could have been beneficial at some historical stages.

LDL becomes immune stimulatory and proinflammatory during oxidation and such oxLDL is abundant in atherosclerotic plaques. Both PCSK9 inhibition and statins immunomodulate this effect where underlying mechanisms are similar but not identical and PCSK9 inhibition may turn out to be relatively strongly anti-inflammatory. PCSK9 may play a role in CVD and autoimmune diseases, by other mechanisms than LDL lowering, and further studies are needed to elucidate if they could be of use in diseases as SLE, where OxLDL, CVD, atherosclerosis, and inflammation are all increased. It would not be the first time in medical history that a drug has been developed, and works, but only partly through the mechanisms the drug development was based.

A weakness in this field, including pleiotropic effects of statins and now also PCSK9 inhibition, is that it is difficult to discern how big the contribution of anti-inflammatory and immune modulatory effects really is, in addition to LDL lowering. The effects may also differ in importance depending on disease stage and setting. Further research about underlying mechanisms is thus needed.

The ultimate goal in this field is to further understand if PCSK9 inhibition could be used for the treatment and prevention of not only CVD but also inflammatory and autoimmune conditions. Further basic studies of more detailed mechanisms by which PCSK9 may be involved are also warranted. If PCSK9 is proven to play a role in the inflammatory component of atherosclerosis and in inflammatory and autoimmune conditions, the implications could be huge. There is currently no anti-inflammatory treatment available in atherosclerosis and CVD, which has been developed for this purpose. Since atherosclerosis is an inflammatory condition, this obviously is a huge unmet medical need, as CVD is a dominating cause of morbidity and mortality. Likewise, the role of statins in this context, including combinations with PCSK9 inhibition, also deserves further studies. In rheumatic disease, there has been a huge development with biologics, which has changed the course of these diseases. In RA, a problem would be to show potential efficacy above the effects already seen by current medications. In SLE, the therapeutic options are fewer, and it may be that PCSK9 inhibition could be tested first in SLE, one reason being that CVD is so common.

Highlights

• PCSK9 targets the LDL receptor LDLR, which leads to increased LDL levels due to lower expression of LDLR.

• PCSK9 levels are associated with risk of CVD independent of LDL levels.

• PCSK9 is associated with rheumatic diseases as RA and SLE.

• Oxidized LDL is involved in atherosclerosis and CVD and raised in SLE and other rheumatic conditions.

• Oxidized LDL-induced immune activation is ameliorated by PCSK9 inhibition.

• Also other anti-inflammatory effects of PCSK9 inhibition have been described.

Declaration of interest

J Frostegård declares an unconditional, investigator-driven grant from Amgen. The author has no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Additional information

Funding

This study was supported by the Swedish Association Against Rheumatism, the Swedish Heart Lung Foundation, EU (IMI; Preciseads), King Gustav V:s 80-year fund, and a previous unconditional grant from Amgen.