Re-visiting Humoral Constitutive Antibacterial Heterogeneity in Bloodstream Infections

1. Inter-individual heterogeneity in serum antibacterial factors of constitutive innate immunity

The antibacterial ability of normal human serum has been observed since the 19^th century,¹⁰ yet studies to assess or understand its inter-individual heterogeneity have been surprisingly limited.^15–17 From a physiological perspective, variability in intrinsic factors such as age, gender, co-morbidities, genetics or epigenetics, or heterogeneity of the blood compositions can have an impact on immune response to pathogen challenge.^18,19 Despite the myriad of potential factors, with decades of intensive studies on antibody repertoire,²⁰ the prevailing notion behind inter-individual heterogeneity in serum antibacterial response is primarily based on the differential antibody repertoire derived from germline, vaccination, and past exposures, activating complement via the classical pathway to rapidly clear bacteria through the formation of trans-membrane pores.^10,21–24 However, other antibody-related factors, including variability in maintaining serum antibody levels from decay²⁵ or efficiency in complement activation by different immunoglobulin subclasses, may also contribute to inter-individual differences in antibacterial activity.^10,26–28 Some antibodies can even “cloak” the bacterial membrane from complement mediated killing. For example, patients with bronchiectasis²⁹ and urosepsis³⁰ were shown to have anti-LPS IgG2 antibody which can paradoxically suppress the complement against Pseudomonas aeruginosa and Escherichia coli.

Prototypical humoral PRMs, such as mannose-binding lectin (MBL) or pentraxins (e.g. C reactive proteins, serum amyloid P component, pentraxin-3), are critical components of constitutive host defense that exert variable complement-dependent antimicrobial effects. Humans have seven common haplotypes that regulate the constitutive serum levels of MBL resulting in stratification into high, intermediate, or low levels and a combination of structural mutations and promoter haplotypes results in MBL deficiency in about 25% of individuals.^31–35 Subsequently, these thresholds of MBL levels have been associated with increased mortality or severity of infection. For example, cystic fibrosis patients with low-secreter MBL phenotype have a reduction in life-expectancy of 5–8 years.³⁶ However, under certain circumstances, low MBL haplotype may instead protect the host against certain infections or hyperinflammatory responses. Low MBL levels appear to be favorable in controlling the development of visceral leishmaniasis in vivo,³⁷ and MBL null mice had enhanced survival in an acute septic peritonitis model, which was correlated with a muted proinflammatory response.³⁸ Similarly, pentraxins possess both antibacterial and immunomodulatory functions. Although they are generally considered as inducible acute phase proteins, some are detectable in variable basal levels in serum of healthy populations and their increased levels have been associated with sepsis and its complications.³⁹ Two intronic SNPs in the pentraxin-3 gene have been associated with increased susceptibility to pulmonary tuberculosis, cystic fibrosis, and urinary tract infections.^40–42 Besides antibodies and the complement-related mechanisms, serum factors such as plakin, lysozyme, β-lysin, and histones have been known to have antibacterial activity,¹⁰ and recent discoveries have showcased the antimicrobial potential of other components in serum (Table 1).⁴

However, these haven’t been extensively studied for their contribution to inter-individual heterogeneity. For instance, peptidoglycan recognition proteins (PGLYRPs) are soluble secreted proteins that can bind to multiple bacterial components and induce exaggerated stress responses resulting in bacterial killing.⁶⁰ In particular, the function of PGLYRP2, found in serum, can be enhanced differentially by other serum components that can affect permeabilization of Gram-negatives’ external membrane.⁶¹ Another example is antimicrobial peptides (AMPs) and proteins. Cathelicidins and defensins are two representative AMPs in humans that bind directly to negatively charged bacterial membranes to disrupt membrane integrity.⁶² While six variants of α defensins in humans are well studied, most variants of β defensins have not been deeply characterized yet.⁶³ Furthermore, there is a void in the field with regards to AMPs unique to serum. Generation or degradation of AMPs in serum can change based on the host conditions such as underlying diseases^64,65 or molecular characteristics of produced AMPs.⁶⁶ These acquired or innate factors may affect the individual repertories of AMPs in serum. Another humoral innate antibacterial mechanism is mediated through nutritional depletion of pathogens. One example is serum lactoferrin, a highly cationic glycoprotein that has antibacterial activity, partly through binding and sequestering iron from being accessible to pathogens.⁵⁷ Lactoferrin also has direct immunomodulatory effect that bridges the innate and adaptive responses. Hepcidin, haptoglobin and hemopexin have also been shown to prevent utilization of circulating free iron by pathogens.³⁵ For all the aforementioned serum factors, it remains unclear whether they are expressed in either or both constitutive and inducible manners, their antibacterial spectrum, and the extent of their contribution to inter-individual heterogeneity. There are additional components in serum, such as cell free DNAs, RNAs, phageome or exposome, with antibacterial activity which may also influence inter-individual heterogeneity but may not be considered as part of the host immunity.^{12–14,58,59} The inter-individual heterogeneity in host innate humoral response is likely multi-factorial. But despite molecular diversity involved, given that their constitutive expression consumes energy to remain operative, we surmise that humoral innate immune factors likely exert their antimicrobial effects through a limited number of shared common mechanisms, including complement activation, agglutination, or neutralization.

Compounding to this inherent complexity, invading bacterial strains have also evolved a myriad of mechanisms to evade the host defenses. Host defenses typically target bacterial cell surface structures like capsules, outer membrane proteins, and lipopolysaccharides, which are not conserved or identical even across all the strains within a single species.^67–70 Bacteria are known to produce strain-specific extracellular polysaccharide capsules that protect against humoral factors and express complement inhibitors.⁷¹ Another example is the incorporation of host cell protectin (CD59) into the bacterial cell membrane has been shown to inhibit trans-membrane pore formation thereby becoming resistant to complement.⁷² Given the complex host-pathogen interplay, it is likely that the varied mechanisms accounting for inter-individual heterogeneity may also be pathogen-specific.

Our group is currently studying the inter-individual heterogeneity of serum antibacterial activity. We found that inhibiting complement did not completely abrogate antibacterial activity and there was a lack of correlation between bacterial specific antibody titers and the extent of bacterial killing by serum from healthy individuals. This heterogeneity is likely caused by humoral innate immunity rather than adaptive immunity, stirring the question whether heterogeneity in humoral innate immunity is inherited or trained. Immunological memory, a hallmark of the adaptive immune response, has been suggested in innate immune cells. This so called “trained immunity”, realized by epigenetic reprogramming of monocytes/macrophages or natural killer cells, is less specific and of shorter duration than adaptive immunity, therefore allowing not only protection against reinfection from the same inciting pathogen but cross-reactive protection against other pathogens as well.^73,74 Interestingly, a recent study on innate immune memory of an invertebrate model, known to lack adaptive immunity, observed that after a primo-infection the innate immune memory can elicit an exclusive host defense through humoral factors (e.g. C-type lectins) without any observable cellular immune response.⁷⁵ Assuming immunological memory is an evolutionary conserved phenomenon, it is conceivable that humoral factors in human serum may also play a significant role in trained innate immunity. As such, the humoral innate “immunobiography”, owing to exposure history, immune memory, and its plasticity, can also contribute to inter-individual heterogeneity.⁷⁶

2. Potential Clinical Utility of Measuring Pathogen-specific Host Serum Antibacterial Activity

Immune functional assays (IFAs) that record the response to ex vivo immune challenges can improve our understanding of the immune system and stratify patient’s immune status for precise diagnosis, prognosis, and therapeutic decision-making.⁷⁷ Unlike with the use of select biomarkers limited to assessing components within a complex immune system, IFAs can provide an unbiased holistic functional assessment of the immune response specific to the stimulant. Serum antibacterial assay, also known as serum bactericidal assay (SBA), is an IFA of both innate and adaptive humoral immune response often used to evaluate the complement-mediated functional activity of infection- or vaccine-induced antibodies. An extensive body of evidence over the past century studying serum antibacterial activity tend to support a positive correlation between serum resistance, which is the bacterial resistance to serum’s antibacterial effect, and the ability to cause bloodstream infections, septic shock, and death—particularly involving Gram-negative organisms.¹⁰ However, a study of patients with Gram-negative bacteremia reported only 37% of the blood isolates were completely resistant and 26% were markedly sensitive to healthy blood serum, suggesting that heterogeneity in both host and microbial factors may play a role.⁷⁸ Most efforts to understand the clinical relevance of serum resistance as a bacterial pathogenicity determinant were based on SBA performed in non- “autologous” format—testing infecting isolates against healthy sera or bacteremic sera against non-infecting strains.¹⁰ To better determine the clinical usefulness of SBA in the context of BSI, we propose two potential use cases based on an autologous testing format, by testing the viability of the infecting bacterial isolate in the same patient’s serum, for predicting or diagnosing bacterial BSI.

Extravascular infections can often translocate into the vascular system and can result in BSIs with an incidence rate of 2 million cases annually in North America and Europe.⁷⁹ UTI is the most common cause of bacteremia, known as urosepsis, with a mortality rate as high as 33%.^80–82 Studies have reported rates of 15–32% for bacteremia amongst emergency department or hospitalized patients with UTI.^83–85 In the elderly population, 40–57% of bacteremia are of urinary origin.⁸⁵ Unfortunately, no current laboratory test or prognostic risk scores can reliably predict the risk of developing urosepsis. Uropathogenic Escherichia coli (UPEC) isolated from patients with pyelonephritis or bacteremia have been shown to exhibit much higher serum resistance (82–93% and 72–80%, respectively) than fecal E. coli isolates (57%),⁸⁶ and bacterial factors that enable serum resistance in these strains are key virulence traits for the development of urosepsis.^87,88 Recently, Coggin et al found that 24% of patients with UPEC-mediated sepsis expressed “inhibitory antibodies” capable of preventing serum-mediated killing of their infecting isolates, highlighting the importance and heterogeneity of host humoral factors in the pathogenesis of urosepsis.³⁰ Autologous SBA performed on infecting isolates from patients with UTI may help predict individual risk of urosepsis and inform appropriate antimicrobial treatment.

Another potential clinical use case of SBA is determining the causality of extravascular isolate in blood culture negative sepsis. While the risk of mortality increases with severity of sepsis, only less than 13% of the patients with presumed sepsis have positive blood culture results despite efforts to reduce false negatives.^89,90 Concurrent extravascular bacterial infections are often detected in blood culture negative sepsis, up to 96% was reported in one ICU study with some patients having more than one site of infection.⁹¹ Given recent recommendations to promote antimicrobial de-escalation in blood culture negative sepsis, determining the causality of extravascular isolates based on autologous SBA to assess their sensitivity to patient’s serum can potentially provide supportive data to better guide appropriate antimicrobial therapy, additional treatment modalities if needed, or the need to search for other sources of infection.^92,93 With improved assay standardization and the use of advanced microfluidic technologies for rapid automated microbial isolation from extravascular infections, autologous SBA may be performed on saved serum prior to antimicrobial initiation for more timely reliable results.

3. Conclusions and future research

As we have described, despite long-standing recognition of serum bactericidal activity, much remains unclear about the mechanisms underlying the inter-individual heterogeneity of this immune functional characteristic and its clinical implications in the context of BSI and sepsis. With advances in translational omics-based technologies (e.g., genomics, transcriptomics, proteomics, metabolomics, or combination thereof), we now have the tools to deconvolute the molecular basis of this heterogeneity in a comprehensive manner and potentially gain novel insights into constitutive humoral innate immunity and trained immunity. To better ascertain the clinical relevance of SBA combined with rapid isolation of infecting bacteria directly from clinical sample would require more large-scale studies that account for pathogen-specific inter-individual heterogeneity by testing in an autologous format. Given the current emphasis to subgroup sepsis patients into endotypes based on underlying immunobiology to better address the heterogeneity in sepsis syndrome, SBA may prove useful to define a distinct sepsis endotype, enabling more precise prediction of disease severity, risk stratification, and opportunities for individualized treatment. Hence, combining emergent technologies to enhance the performance of decades old SBA holds promise in propelling the clinical utility of SBA “back to the future” towards personalized BSI and sepsis management.