Cross-Reactivity as a Mechanism Linking Infections to Stroke

This is way out of my league.

Cross-Reactivity as a Mechanism Linking Infections to Stroke

Guglielmo Lucchese^1,2*, Agnes Flöel¹ and Benjamin Stahl^1,3,4,5

• ¹Department of Neurology, University of Greifswald, Greifswald, Germany
• ²Department of Computing, Goldsmiths, University of London, London, United Kingdom
• ³Department of Neurology, Charité Universitätsmedizin Berlin, Berlin, Germany
• ⁴Department of Neurophysics, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany
• ⁵Psychologische Hochschule Berlin, Berlin, Germany

The relevance of infections as risk factor for cerebrovascular disease is being increasingly recognized. Nonetheless, the pathogenic link between the two entities remains poorly understood. Consistent with recent advances in medicine, the present work addresses the hypothesis that infection-induced immune responses may affect human proteins associated with stroke. Applying established procedures in bioinformatics, the pathogen antigens and the human proteins were searched for common sequences using pentapeptides as probes. The resulting data demonstrate massive peptide sharing between infectious pathogens—such as Chlamydia pneumoniae, Streptococcus pneumoniae, Tannerella forsythia, Haemophilus influenzae, Influenza A virus, and Cytomegalovirus—and human proteins related to risk of ischemic and hemorrhagic stroke. Moreover, the shared peptides are also evident in a number of epitopes experimentally proven immunopositive in the human host. The present findings suggest cross-reactivity as a potential mechanistic link between infections and stroke.

Introduction

When considered separately from other cardiovascular diseases, stroke ranks fifth among all causes of death (1) and, critically, its incidence is on the rise (2).

The etiology of stroke is multifactorial with various environmental and genetic risk factors. Hypertension, diabetes and insulin resistance, smoking, dyslipidemia, obesity, heavy alcohol consumption, atrial fibrillation, and carotid stenosis are all established and well-investigated modifiable risk factors of stroke (3–5).

Additionally, there is evidence that environmental factors may also increase risk of stroke, including viral and bacterial infections, such as periodontitis (6) and respiratory infections (7), and infection with Chlamydia pneumoniae (8) or Cytomegalovirus (9). However, relatively little is known so far about the role of different pathogens as well as the molecular basis and the mechanisms that potentially link infections to stroke.

Here we set out to investigate whether or not infections can induce immune responses capable of cross-reacting with human proteins that, when altered, have been associated with stroke. Our hypothesis was that immune responses induced by infectious agents might cross-react with crucial stroke-related proteins, thus contributing to the multifactorial pathogenesis of cerebrovascular disease.

To address this hypothesis, we analyzed pathogens, as well as proteins that are known to be associated with increased risk of ischemic and hemorrhagic stroke by searching for common peptides that might underlie cross-reactions.

Specifically, we analyzed antigens from the following pathogens that have been reported to have a possible influence on stroke: the periodontal bacterium Tannerella forsythia (10), Haemophilus influenza (11), Streptococcus pneumoniae (7), Chlamydia pneumoniae (8), Influenza A viruses (12, 13), and Human Cytomegalovirus (9).

Methods

We analyzed the amino acid (aa) primary sequence of pathogen antigens (with short name and UniProt ID in parentheses):

• Surface antigen repeat/outer membrane protein (OMP; UniProtKB: A0A0F7WYE8_CHLPN) from Chlamydia pneumoniae;

• Pneumococcal vaccine antigen A (PVAA;UniProtKB: PVAA_STRR6) from Streptococcus pneumoniae;

• Surface antigen BspA (BspA; UniProtKB: O68831_TANFO) from Tannerella forsythia;

• Outer membrane antigenic lipoprotein B (LPPB; UniProtKB: LPPB_HAEIN) from Haemophilus influenzae (strain ATCC 51907);

• Hemagglutinin (HA H1N1; UniProtKB: HEMA_I34A1) from Influenza A virus (strain A/Puerto Rico/8/1934 H1N1);

• Hemagglutinin (HA H5N1; UniProtKB: HEMA_I96A0) from Influenza A virus (strain A/Goose/Guangdong/1/1996 H5N1);

• Hemagglutinin (HA H3N2; UniProtKB: HEMA_I68A6) from Influenza A virus (strain A/Northern Territory/60/1968 H3N2); and

• 65 kDa phosphoprotein (pp65; UniProtKB: PP65_HCMVM) from Human Cytomegalovirus (HCMV; strain Merlin).

The primary sequence of pathogen antigens was dissected into partially overlapping pentapeptides with a one-residue-offset: i.e., MFKRI, FKRIR, KRIRR, and so on. Then, each pentapeptide was analyzed for occurrences within a library consisting of primary sequences of human proteins involved in stroke. The human protein library was a priori chosen from the UniProtKB Database (https://www.uniprot.org) (14) using the keyword “stroke.” We obtained an unbiased list of 74 human proteins (in)directly associated with stroke (Table S1). Stroke-related proteins are indicated as UniProtKB entry names throughout the present article, except when discussed in detail. The pathogen antigens and the human proteins were searched for common sequences using the pentapeptide as a probe unit because a pentapeptide is an immunobiological determinant sufficient for epitope-paratope interaction and for inducing specific immune responses (15–18).

The immunologic potential of the shared peptides was analyzed using the Immune Epitope Database (IEDB; www.iedb.org) (19). All evaluations were based only on epitopic sequences that had been experimentally validated as immunopositive in the human host.

This linear peptide similarity analysis procedure has been used and described before (20, 21).

Results

In a detailed overview, Table 1 shows that 49 out of the 74 human stroke-related proteins share peptide sequences with antigens from pathogens that proved to be (in)directly involved in stroke (6–10). It can be seen that

• The pathogen vs. human peptide overlap is unexpectedly high when considering that the probability for two proteins to share a pentapeptide is 1 out of 20⁻⁵, that is, 0.0000003125 or close to zero.

• The peptide overlap varies widely, with T. forsythia BspA and Influenza A HA H3N2 being the pathogen more and less involved in the peptide sharing, respectively.

• The high number of stroke-related proteins involved in the viral peptide overlap precludes a detailed protein-by-protein analysis. However, an example worth noting is the human ATP-binding cassette sub-family C member nine (ABCC9 or SUR2) that shares peptide sequences with all of the pathogen antigens analyzed, with the exception of the Influenza A HA H3N2 virus. ABCC9 is a subunit of ATP-sensitive potassium channels (K_ATP) that can form cardiac and smooth muscle-type KATP channels with KCNJ11 and mediates neuroprotection (22).