Antibodies against phosphorylcholine and protection against atherosclerosis, cardiovascular disease and chronic inflammation

ABSTRACT

Introduction

Chronic inflammatory diseases include cardiovascular disease (CVD), atherosclerosis, rheumatic and autoimmune diseases, and others, constituting a large part of the disease burden. It is therefore of major importance to improve understanding of underlying mechanisms, prediction, and treatment.

Areas covered

Broad fields including atherosclerosis, immunology, and inflammation are covered, through searches on PubMed and background knowledge. Phosphorylcholine (PC) is both a danger-associated molecular pattern (DAMP), present on oxidized LDL (OxLDL) in atherosclerotic lesions and dead cells, and a pathogen associated molecular pattern (PAMP), present on microorganisms. IgM and IgG1 antibodies against PC (anti-PC) are associated with protection in several chronic inflammatory conditions, especially in CVD and atherosclerosis where most research has been done. PC-immunization ameliorates atherosclerosis in animal models and several potential underlying mechanisms have been proposed, including anti-inflammatory, decreased uptake of OxLDL in the artery wall, promotion of T regulatory cells. Anti-PC develops during the first years of life. Low levels of IgM and IgG1 anti-PC may be caused by lack of exposure to microorganisms, including nematodes and helminths among others.

Expert opinion

anti-PC could improve prediction of clinical outcome and raising anti-PC could be developed into a novel therapy.

KEYWORDS:

• Antibodies
• atherosclerosis
• immunity
• inflammation
• phosphorylcholine

1. Introduction

Chronic inflammation is a common theme in several different diseases which together represent a huge part of disease burden today. The largest of these in terms of mortality is cardiovascular disease (CVD), including myocardial infarction (MI) and stroke. The major underlying factor is atherosclerosis, which is the main topic herein, not only because CVD and atherosclerosis are dominant as disease causes, but also because much more is known about phosphorylcholine (PC) and antibodies against PC (anti-PC) in these conditions as compared to other chronic inflammatory conditions. Among these, rheumatic diseases are typically of inflammatory and also chronic inflammatory nature and nowadays also osteoarthritis is usually seen as an inflammatory and rheumatic condition. In addition, type 2 diabetes and insulin resistance and even dementia and obesity are examples of chronic inflammation[1].

Traditional risk factors for CVD (and atherosclerosis) include hypertension, diabetes, smoking and hyperlipidemia, e.g. high cholesterol and especially high low density lipoprotein (LDL). In pioneering studies of how LDL could become atherogenic, oxidized LDL (OxLDL) was implicated [2] and many studies have demonstrated different mechanisms by which such OxLDL could contribute to development of atherosclerosis and ensuing CVD. These include proinflammatory effects on endothelial cells and macrophages, activation of T cells, promotion of cells death and being taken up by macrophages and other cell types which develop into inert foam cells[1].

Atherosclerosis is characterized by accumulation of dead cells, oxidized phospholipids, especially present in oxidized low density lipoprotein (OxLDL) and activated immune competent cells which produce mainly proinflammatory cytokines, such as TNF, IL-1 and IL-6 and chemokines among other substances. In addition, calcification is another important feature of atherosclerosis, most likely being a consequence of inflammation and accumulation of dead cells [1,3]. The lack of clearance of dead cells is likely to be a major factor in atherosclerosis development, since accumulating cell debris can contribute to inflammation in the plaques[1].

2. PC and related antigens

The cellular stress and inflammation in atherosclerotic lesion are characterized by generation of damage associated molecular patterns (DAMPs) which are expressed on both damaged and dead cells and on oxLDL. There are several different types and categories of DAMPs, some of which which are proinflammatory and immune stimulatory. These latter include malondialdehyde (MDA) and phosphorylcholine (PC)[1]. OxLDL is proinflammatory and immune stimulatory [4,5] and we reported that PC is one possible cause of this, promoting induction of immune activation and production of interferon-gamma[6].

Another example is cardiolipin (CL) and especially oxidized CL (OxCL) which are present in mitochondria, and bacteria (mitochondria believed to be derivatives of bacteria from an evolutionary point of view). OxCL has proinflammatory properties [7] and both OxLDL:s and OxCL:s proinflammatory effects are inhibited by Annexin A5, a protein with anti-inflammatory properties [7,8].

PC is the polar headgroup of phosphatidylcholine, which is an abundant membrane phospholipid[9]. Enzymatic and oxidative modification of phosphatidylcholine in cells and LDL can be oxidatively modified, both through enzymes such as phospholipase A2 (PLA2) and by chemically induced oxidation. Of note, when oxidized phosphatidycholine loses fatty acids in the sn-2 position which gives bioactive properties to such oxidized phospholipids (OxPL) which thus contain PC in an exposed position[10]. Interestingly, while PC is abundant and exposed in atherosclerotic plaques on dead cells and OxLDL, in human plasma a major carrier is lipoprotein (a) (Lp(a)[11], which can induce arterial inflammation[12]. Still, also PC on oxLDL is present in plasma in the circulation and is raised in individuals with high risk to develop CVD, including hypertensives and patients with systemic lupus erythematosus (SLE) [13–15].

Such PC-exposing lipids have several interesting properties, participating in and promoting apoptosis induction, endothelial dysfunction and also endoplasmatic reticulum (ER)-stress and mediate vascular inflammation [9,16,17]. In a mouse model, oxidized phospholipids with PC as one mediator were demonstrated to be atherogenic and proinflammatory[18].

There is an interesting difference between epitopes generated during LDL-oxidation and other events leading to DAMP-exposure, namely that PC is also a pathogen associated molecular pattern (PAMP), which is not the case with MDA, at least not as is presently known. One well-known example is the cell wall of Streptococcus pneumoniae. Interestingly, there is a variability in how effective such anti-PC against carbohydrate-bound PC is in its effects, both against the bacteria and in relation to other disease processes [19–22]. Of note, immunization with S pneumoniae confers some protection against atherosclerosis in mice, though this is relatively modest[23].

Isotypes and subclasses of anti-PC are discussed below, but it is interesting to note that the mode of presentation of PC may play a major role in details of the anti-PC response. Other presentations of PC as PAMP include on proteins, related to nematodes, helminths and worms. We reported that infection with Treponema is strongly associated with anti-PC levels at Kitava, Papua New Guinea[24]. PC binds many proteins avidly, and conjugates between PC and albumin and with keyhole limpet hemocyanin (KLH) appears to be very similar as antigens[1]. However, when PC is conjugated with tuftsin, thus not related to inflammatory lipids, and being a small molecule (not inducing antibodies) it may ameliorate inflammatory diseases [25,26].

To the best of my knowledge, there is no evidence that anti-PC binds PC in normal cells. Clearly, more research on the details of the antigenicity of PC is warranted.

3. Antibodies in chronic inflammation

Even though antibodies against LDL modified and/or oxidized through different mechanisms (anti-OxLDL) have been known for a long time, their role and clinical associations remain unclear. In early studies, anti-OxLDL were reported as risk markers, associated with CVD and atherosclerosis [27–31].

To the best of my knowledge, the first time antibodies were reported to be associated with protection in the context of CVD is a study where we used two different methods, both yielding results where anti-OxLDL was negatively associated with CVD, in this case borderline hypertension, thus being a protection marker in humans[32].

Other reports are in line with this finding, where anti-OxLDL is a protection marker [33–35].

Taken together, anti-OxLDL has been reported to be both a risk and a protection marker for CVD and atherosclerosis [1,27,32,34].

These discrepancies could be related to the complexity of the LDL-oxidation process where antigens exposed could vary to some extent, and there may also be differences in methods used for antibody determination which play a role[1].

It is also possible that anti-OxLDL could form immune complexes which could be of clinical relevance, and may be risk markers and even cause inflammation[36]. One example of immune complexes containing OxLDL being associated with risk is in SLE, where complexes with beta-2-glycoprotein I, an important antigen for thrombotic anti-cardiolipin antibodies[37].

An interesting possibility is that some anti-OxLDL antibodies cross react with anti-CL which are thrombogenic, especially in SLE and the antiphospholipid antibody syndrome[38].

There are other interesting and immunogenic antigens related to OxLDL. One is the protein in LDL, apoB100 which can also be modified with eg MDA. However, antibodies against apoB or its peptides appear to give varying results, like anti-OxLDL and in some cases associated with protection and in others with worse outcome [39–41].

Another interesting antigen generated during LDL-oxidation is malondialdehyde (MDA) and MDA can be conjugated with carrier proteins. IgM anti-MDA is a protection marker in CVD or related conditions with high risk as CVD, such as SLE and uremia, though they have been less studied in humans than IgM anti-PC [42,43].

4. Anti-PC as a risk marker in chronic inflammation

We reported that IgM anti-PC is negatively associated with atherosclerosis progress among hypertensives[34]. Based on this, the hypothesis that low IgM anti-PC is a novel risk marker for atherosclerosis and CVD was proposed[44]. We published further reports indicating IgM anti-PC as a marker for protection both in CVD and atherosclerosis [45–51]. In follow up publications we reported that IgM anti-PC is associated with protection in CVD (including both stroke and MI), and with fast carotid intima media thickness progression and cardiovascular risk in men[45].

Other studies have confirmed the notion that IgM anti-PC are protection markers in atherosclerosis/CVD [34,52–56], and also in venous thrombotic disease[57]. However, our finding that patients with acute coronary syndrome (ACS) and low IgM anti-PC have a worse prognosis than those with high levels [49], was not confirmed in another study [58]. In our study, the associations were weak, since they were only present significantly at the lowest levels, and it is possible that in an acute situation, other factors such as ongoing inflammation, and even consumption of anti-PC could play a role.

Of note, IgM anti-PC constitute an important part of the circulating IgM pool, about 5–10% has been reported. Given the role of IgM anti-PC as a protection marker in different contexts, the properties of other subclasses and isotypes could give of importance, one reason being the potential for immunization with PC to ameliorate chronic inflammatory diseases[1].

Both SLE and chronic kidney disease (CKD) and uremia are interesting examples of diseases where the risk of CVD and atherosclerosis is very high. These conditions could be used as a kind of ‘human model’ of immunity and atherosclerosis in addition to being an important clinical problem.

The risk of CVD is reported to be 50 times increased in SLE [59]. A combination of traditional and non-traditional risk factors and also associated with atherosclerosis can provide an explanation for this [15]. In line with this, atherosclerosis is increased in SLE as compared to age- and sex matched controls which is the case especially for prevalence of atherosclerotic plaques is increased [15,60,61]. Since fissures and rupture of plaques are likely to be of major importance in CVD [1] we determined that echolucent plaques, which may indicate vulnerable atherosclerotic lesions and their prevalence is increased in SLE [62–64]. A fully human IgG1 anti-PC clone increases clearance of dead cells and that IgG1 but not IgG2 anti-PC is associated atheroprotection in SLE[65].

IgM and IgG1, more than IgG2 anti-PC has protective properties associated with less atherosclerosis progress, vulnerable plaques and mortality in CKD [66–68], and also with less vascular calcification, an important aspect of atherosclerosis, in CKD[69].

Also associations in other disease conditions including rheumatic have been reported. In line with its role in CVD and atherosclerosis, IgM anti-PC is associated with protection against CVD in Rheumatoid arthritis[70].

IgM anti-PC is a protection marker for CVD in Rheumatoid arthritis (RA) [70] and low levels of anti-PC are associated with prevalence of atherosclerotic plaques after 5 years [71], and also associated with being a non-responder to biologics (TNF-inhibitors) in RA[72]., in both osteoarthritis [73] and vasculitis[74]. Furthermore, IgM anti-PC is associated with protection against upper respiratory infections[75].

IgA anti-PC appears to be more complex, than other isotypes and subclasses. We reported that IgA anti-PC is a protection marker for atherosclerosis [66] while another study (using PC conjugated with KLH as an antigen) reported that IgA was associated with long-term risk of CVD as a risk marker (in contrast to IgM anti-PC in that study)[76]. IgA in general is an important part of the humoral immune system at mucosal surfaces, and its role in the circulation is not believed to be of major importance. More research on the role of IgA in chronic inflammation is warranted.

In previous studies we were not able to determine IgG3 and IgG4 anti-PC at significant levels and are not aware of other studies demonstrating this[66].

The observation that IgG1 anti-PC is more associated with protection than IgG2 anti-PC could be of interest both mechanistically and for prediction and potential therapy. PC can be presented directly bound to a carrier, e.g. lipid or protein, but also as p-nitrophenyl phosphorylcholine (NPPC) and since long, it has been known that human anti-PC can be divided into group I (IgM and IgG1) and group II (IgG2)[77].

While group II antibodies only recognize NPCC, where also the phenyl-ring attached to PC is antigenic, group I anti-PC recognizes both forms of PC. IgG2 anti-PC recognizes capsulated bacteria, with carbohydrate antigens, and also bactericidal properties [66,78,79]. This could provide an explanation to why the risk of CVD is raised in periodontitis [80], in spite of increased IgG2 anti-PC levels [81,82].

Of the five Ig isotypes, IgM, IgG, IgA, IgD and IgE, the first three have been studied, and IgM anti-PC has most documented protective properties. We have not been able to determine anti-PC IgG3 and IgG4 among the four subclasses, but it is interesting to note that IgG1 and IgG2 in general differ in their properties, which may provide an explanation as to why IgG1 but not to the same extent IgG2 anti-PC is associated with protection in different contexts. While both participate in complement activation, only IgG1 in general is associated with binding to FC receptors, a mechanism which is important phagocytosis, including clearance of dead cells by macrophages and other cells. If I may speculate, this may indicate that clearance of dead cells could be a major factor in both SLE and atherosclerosis. However, also antigen presentation is an effector function of IgG1, and further research is needed to determine the properties of different subclasses of IgG anti-PC even though bacterially induced anti-PC thus is associated with less protection. An intriguing possibility is that PC presented by other pathogens, eg nematodes and parasites, promotes more protection[83].

I am not aware of any publications where IgM or IgG1 anti-PC has been implicated as a risk marker. Also, we have not seen an increased risk in individuals with very high IgM (or IgG1 antibodies) in any studies, including SLE.

Interestingly, PC binding myeloma proteins, e.g. anti-PC were identified already in the mid-70s.[84] Anti-PC has traditionally been described as natural antibodies, which are germ line encoded, and present from birth and where one clone, TI5 is dominant. This information has largely been obtained from laboratory mice. However, we could not reproduce this in human systems. When we produced fully human anti-PC clones, there was no dominant among these. Instead, human anti-PC appears to be somatically mutated and have undergone somatic mutation with Ig -switch. Furthermore, and in line with this, they are T cell dependent [42,85].

We also reported that humans are born with very low levels of anti-PC, in sharp contrast to anti-MDA. Not even during the first two years of life do humans come even close to mothers´ levels. Most likely, the environment therefore plays a role, including microbiome, but very little is known about underlying mechanisms and microorganisms[86]. In line with this, the genetic contribution of IgM anti-PC is relatively modest[87].

5. Experiences from the animal kingdom

Animal models of atherosclerosis have also given important information about the role of inflammation and immunity. An inherent problem with the mouse models usually used is that mice – or rats which are sometimes also used – do not develop atherosclerosis naturally, not even under laboratory conditions, if not genetically manipulated. Also rabbits have been used, but these also have similar problems. The most commonly used mice models have very high blood lipids, so it can be argued that the atherosclerosis-like disease they develop is only reflecting some of the human atherosclerosis, and they do not seem to develop CVD as humans.

Still these models can be very useful, also to simulate a more acute event with interventions, e g cuffs.

An important milestone in studies of immunity and atherosclerosis is the finding that immunization can modulate atherosclerosis using heat shock protein (HSP)60/65 as an antigen. In this case, atherosclerosis was increased[88]. Also immunization with OxLDL leads to less atherosclerosis [89,90].

In mice models, an atheroprotective role of anti-PC is suggested both using active [91] and passive [92] immunization which is the case also SLE[93], and RA[94]. Also Pneumococcal immunization leads to a modest but significant decrease of atherosclerosis, and increased levels of different antibodies including anti-PC[23].

One of the few animals which develop atherosclerosis under natural conditions is the African elephant where signs of the disease conditions were described in studies in the 60s.[95] In contrast, in captivity and as domestic animals, there are many examples of atherosclerosis development[1].

Interestingly, brown bears (Ursus arctos) do not develop any signs of atherosclerosis, in spite of immobilization during hibernation lasting 5–6 months during winter, strong weight gain, kidney insufficiency high blood lipids. In line with the hypothesis herein, they develop strikingly high levels of IgM and IgG1 anti-PC and we proposed that this may represent a potential natural immunization against atherosclerosis[96].

6. Potential underlying mechanisms

Several experimental mechanistic studies in addition to animal studies support the notion that anti-PC is not only a protection marker, but could also play a causative role, protecting against atherosclerosis, CVD, and other types of chronic inflammatory conditions. Also here, most studies have been performed using IgM anti-PC, but IgG anti-PC inhibits pro-inflammatory effects on endothelial cells caused by PC-exposing oxidized lipids[51].

IgM anti-PC has other anti-inflammatory, immunomodulatory properties, ameliorating proinflammatory immune activation, by promoting polarization of T cells to T regulatory cells. This was demonstrated in ex vivo systems using cells from healthy donors, atherosclerotic plaques, and SLE-patients[97].

Most likely, build up of OxLDL in macrophages, which develop into inert foam cells, is an important part of atherosclerosis development. Most likely, IgM anti-PC induced inhibition of macrophage-uptake of oxLDL is atheroprotective[48]. Accumulation of dead cells is a major component of atherosclerosis, and IgM anti-PC-induced inhibition of cell death caused lysophosphatidylcholine, generated during LDL-oxidation and toxic, could be yet another protective property[66]. It is possible that the most important atheroprotective effect of IgM anti-PC is its ability to increase clearance of dead cells[42], which is a property also of IgG1 anti-PC[65]. Of note, this may be of major importance also in SLE, where defective clearance of dead cells may be a major feature in addition to increased expression of PC in LDL[1].

7. A development of the old Friends/Hygiene hypothesis: causes of low anti-PC

Both CVD and chronic inflammatory conditions are more or less absent in Kitava, Papua New Guinea[98]. This study population could shed light on the chronic inflammatory diseases in Western populations and increasingly in the whole world[1].

We reported that, IgG and IgA anti-PC IgM are decreased among Swedes compared with age and sex-matched individuals from this study population from Kitava, Papua New Guinea. Based on this we proposed that this could be one explanation as to why chronic inflammation was very unusual there, in the early 90s when the samples where collected and cohort study created. Most likely, the high anti-PC levels could be related to the infectious panorama since many microorganisms expose PC [99,100].

One possible contributing cause could be exposure to Treponema infection, leading to Jaws. There could be several other such agents, though[24].

PC exposed on proteins in microorganisms as helminths could also be immunomodulatory and prevent inflammation per se [101–103], and 69 proteins with the PC-epitope were identified[103].

Interestingly, ES-62, a nematode protein with huge amounts of PC, ameliorated atherosclerosis and SLE in a mouse model, where anti-PC was raised in parallel, thus in line with other observations immunizing with PC-conjugates[93].

Many such microorganisms, including also parasites, helminths and nematodes have been co-evolving with humans for millions of years, and these could be involved in promoting high anti-PC in Kitava, New Guinea but not in countries like Sweden, where exposure is low[104].

According to the hygiene-hypothesis, the higher prevalence of autoimmune and allergic diseases in Western countries as compared to more traditional societies can to some extent be caused by reduced exposure to environmental agents including infections [105].

In a development of this hypothesis, the ‘Old friends’ hypothesis was proposed. Here focus is on lack of exposure to immune regulatory infectious agents, which previously were common and lack of these could contribute to chronic inflammatory diseases [106–108].

We have thus proposed a more specific development of this hypothesis: anti-inflammatory and protective natural antibodies as anti-PC are too low nowadays in high income countries and this could play a role in autoimmune disease and chronic inflammation, including atherosclerosis and CVD.

8. Conclusions

IgM and IgG1 anti-PC are associated with protection in chronic inflammatory diseases, especially atherosclerosis and SLE, but also in a growing number of related ones. This notion is supported by epidemiological and experimental evidence, where also plausible mechanisms have been described. Immunization to raise these antibodies could therefore be an interesting possibility, which deserves attention.

9. Expert opinion

Even though the role of inflammation in atherosclerosis, and thus CVD, has been known for long time, the focus of research in this field was for rather long time on other aspects, albeit important, such as dyslipidemia. This has led to development of improved prevention and/ treatment of the major established risk factors: hypertension, smoking, diabetes and hypercholesterolemia. Treatment of hypertension with different medications, is arguably one of the real successes of medicine today. Also treatment of diabetes has improved, and hypercholesterolemia can now be controlled by both statins and Proprotein Convertase Subtilisin/Kexin Type 9 (PCSK9)-inhibition. The role of smoking has been highlighted, and incidence of smoking has decreased in many countries. Still, there is no medication available which was developed to decrease the inflammation in atherosclerosis. In line with this, the proinflammatory immune response occurring in atherosclerosis in general remains to be treated. Even though it is likely that both statins and PCSK9-inhibition not only exert their atheroprotective and CVD-decreasing effects due to lipid lowering, but also because of anti-inflammatory and immune modulatory properties, this is difficult to demonstrate in scientific studies. Still, several lines of evidence argue that such pleiotropic effects play a role. In spite of this progress, atherosclerosis and CVD still represent the leading cause of death, not only as previously, in the ‘developed world’ but increasingly in the whole world.

It is therefore of major interest to develop novel preventive and therapeutic strategies. In Rheumatology, the development of biologics, especially different drugs against tumor necrosis factor (TNF) have been a substantial improvement, and it is possible that similar development can occur in atherosclerosis and CVD, albeit likely against other compounds than TNF. Another approach could be immunization, passive or active. Here OxLDL may be one focus, and especially defined parts of it. The key findings in relation to PC is that this small hapten-like compound is exposed on oxLDL, dead cells, and some pathogens, not least nematodes, parasites and some bacteria. PC is both a DAMP and PAMP, and is likely to play a role in chronic inflammation in atherosclerosis, and PC exposed on OxLDL is immune stimulatory and proinflammatory. In relation to anti-PC the key findings are that anti-PC is abundant in humans, and are developed during early life, and is thus not germ line encoded in a way that was suggested from studies on laboratory mice. Anti-PC is associated with protection in CVD and atherosclerosis as reported in many studies, often low levels, e g an immune deficient state, is associated with risk, especially for stroke where the associations sometimes are striking. Also in rheumatic and autoimmune conditions, there are associations between anti-PC and protection, not least in SLE which is of interest since the risk of CVD in SLE is strikingly high. Also the prevalence of plaques is increased. Underlying mechanisms have been elucidated, PC-immunization, both passive and active, have a significant effect to ameliorate atherosclerosis, and clinical associations have been reported by several different groups. The knowledge about how anti-PC is regulated in humans, and how it can optimally be raised needs to be improved. Determination of anti-PC levels could help improve prediction and thus optimize available treatment. Potentially PC-immunization could be combined into a theranostic concept, where individuals at increased risk, and/or with low anti-PC levels (IgG1 and IgM) could be eligible for immunization.

Article highlights

• Chronic inflammation including atherosclerosis and its consequence cardiovascular disease are major health problems in the world

• Phosphorylcholine (PC) is an antigen on dead cells, oxidized lipids and some pathogens

• Antibodies against PC (anti-PC) are abundant and associated with protection against the mentioned disease conditions

• Underlying mechanisms include anti-inflammatory and increased clearance of dead cells

• Raising anti-PC levels through immunization could therefore be developed as a novel therapy

Declaration of interest

J Frostegard is in the board of directors at Annexin Pharmaceuticals and is named as inventor in patents related to Annexin A5. The authors have 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.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was supported by grants from The Swedish Heart Lung Foundation, European Union (Intricare, Preciseads), the Swedish association against Rheumatism, King Gustav V:s 80-year fund.