Potential of Immunomodulatory Agents for Prevention and Treatment of Neonatal Sepsis

Abstract

Prevention of neonatal infection-related mortality represents a significant global challenge particularly in the vulnerable premature population. The increased risk of death from sepsis is likely due to the specific immune deficits found in the neonate as compared to the adult. Stimulation of the neonatal immune system to prevent and or treat infection has been attempted in the past largely without success. In this review, we identify some of the known deficits in the neonatal immune system and their clinical impact, summarize previous attempts at immunomodulation and the outcomes of these interventions, and discuss the potential of novel immunomodulatory therapies to improve neonatal sepsis outcome.

Background and Introduction

Every year, more than one million newborns die from sepsis or serious infection the first four weeks of life¹. Prevention of newborn infection therefore remains a significant global challenge. Neonatal healthcare providers routinely care for infants with risk factors, symptoms, or signs of sepsis, requiring immediate clinical evaluation and presumptive antibiotic coverage. The need for such timely intervention is due to the high risk of mortality in this vulnerable population that is likely the result of multiple immunologic differences between the neonate and older populations. The teleology of these immunological differences is believed to be focused on preventing preterm birth secondary to in utero inflammatory responses, reducing the likelihood of fetal rejection by the mother, and developing fetal immunologic tolerance². While the quiescence of the fetal immune system is likely essential in utero, these responses ex utero may carry disadvantages, including impaired host defense against infection and weak neonatal vaccine responses that can impede efforts to protect this at risk population. Therefore, the newborn is relatively ill-equipped to deal with an infectious challenge and as a result exhibits increased morbidity and mortality compared to children and adults^3,4. These risks are significantly greater in premature, critically-ill neonates with exposure to intrapartum infections, premature rupture of membranes, or in which invasive procedures and mechanical ventilation are required⁵. While numerous studies have identified molecular markers of neonatal sepsis to facilitate early diagnosis, there has been relatively less study of the distinct pathophysiology of neonatal sepsis. Comprehensive investigation into the individual immunological responses of neonatal sepsis may define functional immunodeficiencies that can be targeted to prevent or treat neonatal infection.

Neonatal immune function

Both innate and adaptive immunity are distinct at birth relative to adulthood⁶. Neonatal immune responses are generally T_H2-skewed, being geared towards immune tolerance instead of towards defense from microbial infections [T_H1 skewed] (Table 1)^7–10. Neonatal antigen-presenting cells (APCs) demonstrate impaired production of T_H1-polarizing cytokines (e.g., IL-12, interferon-γ that direct immune responses against microbial pathogens) and impaired up-regulation of co-stimulatory molecules to most Toll-like receptor (TLR) agonists^2,11,12. Nevertheless, certain stimuli such as Bacille Calmette Guerin (BCG) induce adult-like responses via mechanisms that are not yet completely defined^7,13,14. Additionally, neonatal T cells require increased stimulus in order to achieve adult-level responses^7,13,14. Compared to adults, neonates manifest delayed, shortened, and decreased B cell responses that limit their responses to infection and vaccination⁷. Studies in adult mice using models of sepsis indicate that impairments in adaptive immunity result in poor survival^15–17, but the impact of the distinct neonatal adaptive immune system on neonatal sepsis survival has not been defined. Of note, adjuvants directed at decreasing sepsis-induced adaptive immunodeficiencies (e.g., inhibition of T and B-cell apoptosis¹⁸ or enhancement of T cell function¹⁹) improve sepsis outcomes in preclinical adult animal models. However, because the adaptive immune system does not function exactly in the neonate as it does in the adult⁷, we must concede that therapies targeting adult sepsis-specific deficits may not work as effectively in neonates.

Table 1.

Deficits in Neonatal Adaptive Immune Function and the Proposed Clinical Impact.

Adaptive Immune Deficit	Proposed Clinical Impact
Limited antecedent exposure of T cells to foreign antigens	Lack of rapid, strong, memory response
Greater requirement for CD4+ T cell stimulation	Decreased T cell activation, proliferation
TH2skewed and attenuated CD4+ T cell cytokine response	Decreased response to infection, particularly intracellular pathogens
Poor CD4+ T cell-dependent B cell stimulation	Poor antibody production
Decreased CD8+ T cell cytolytic activity	Decreased clearance of intracellularly infected cells
Abundant, potent T regulatory cell population present at birth	Inhibited TH1 T cell responses, decreased response to infection, limit vaccine responses of newborns
Maternal antibodies interfere with B cell antibody response	Attenuated antibody production
Weak humoral response, predominantly IgM	Poor opsonization and clearance of bacteria
Poor antibody response to polysaccharide antigens	Increased susceptibility to encapsulated organisms
Deficient CD40 ligand stimulation of B cells	Poor antibody production-lack of memory response
Underdeveloped spleen and lymph nodes	Poor antibody production, poor clearance of bacteria from blood

The distinct function of the neonatal adaptive immune system renders the neonate especially dependent on the function of the innate immune system and on circulating maternal antibodies transferred during the last trimester, though these are deficient in the setting of prematurity. The neonatal innate immune system is also impaired compared to adults, likely contributing to their increased susceptibility to infection (Table 2)^6,12,20. Neonates have fragile skin (preterm newborns), decreased circulating complement components (limiting opsonization/killing of pathogens), diminished expression of some antimicrobial proteins and peptides (APP), decreased production of type I interferons and T_H1 polarizing cytokines (and a bias towards T_H2-type responses), and quantitative and/or qualitative impairments in neutrophil, monocyte, macrophage, and dendritic cell function¹². Neonatal leukocytes, particular under stress conditions, demonstrate diminished cellular functions necessary for bacterial clearance, including diminished responses to most TLR agonists (constituents of microbes), reduced production of cytokines/chemokines and APP, diapedesis, chemotaxis, phagocytosis, and antigen presentation^6,12. Neonates are also deficient in functional splenic follicules that filter blood and remove pathogens, further limiting bacterial clearance, and increasing the risk of fulminant infection.

Table 2.

Deficits in Neonatal Innate Immune Function and the Proposed Clinical Impact.

Innate Immune Deficit	Proposed Clinical Impact
Fragile, easily disrupted skin (particularly in premature)	Portal of entry for microbes
Decreased serum complement components	Decreased complement-mediated killing and opsonization lead to poor bacterial clearance and decreased naïve B cell activation
Defective neutrophil amplification, mobilization, and function (phagocytosis, respiratory burst, lactoferrin and BPI production)	Poor bacterial clearance
Reduced MHC Class 2 expression on antigen presenting cells (APCs)	Poor T and B cell stimulation
Impaired APC function (decreased TH1 polarizing cytokine production, poor antigen presenting function, impaired mobilization, increased stimulation requirement to effect response)	Poor bacterial clearance
Depressed Natural Killer (NK) cell cytotoxic function	Poor clearance of cells infected with intracellular pathogens
Intrinsic immaturity of dendritic cells (DCs)	Poor antigen presenting function, poor memory response
Impaired cytokine production in response to pathogens	Poor chemotactic gradient formation, poor cellular recruitment to site of inflammation
Decreased neutrophil storage pool in bone marrow	Early depletion associated with poor sepsis outcomes
Decreased opsonin production	Decreased uptake and killing by phagocytes
Impaired response to certain TLR agonists, decreased down-stream signaling following TLR stimulation	Decreased chemotaxis and recruitment of innate cellular defenses

The net effect of these deficits in neonatal immunity is to leave the neonate extremely susceptible to microbial invasion. There is therefore an unmet medical need to prevent and treat neonatal infection, and in this context, immunomodulatory agents are of considerable interest. Can the neonatal immune system be augmented in vivo in order to prevent and/or reduce sepsis mortality?

Attempts at Neonatal Immunomodulation

Experimentally and theoretically sound methods of immunomodulation directed at improving sepsis outcomes by correcting some neonatal immune deficits have been attempted without major improvement (Table 3)^21–33. Therapies that enhance the quantity and/or quality of neutrophils, including granulocyte transfusions, GM-CSF, and G-CSF have been studied in human neonates^34–38. These agents increase circulating numbers of neutrophils and appear to be safe, but have thus far been unsuccessful at significantly reducing neonatal sepsis mortality^21,22. The administration of intravenous immmunogoblulin (IVIG) (polyclonal and monoclonal) administration has been aimed at both prevention and treatment of bacterial infection in neonates^39,40–48, but meta-analyses have demonstrated only a marginal clinical benefit in prevention and there is insufficient data at present to routinely recommend IVIG for treatment of neonatal sepsis²⁶. To more fully address the latter question of whether IVIG reduces sepsis (suspected or proven) mortality, a multi-center, international, double-blind, randomized controlled clinical trial (International Neonatal Immunotherapy Study) has been completed with publication of the results anticipated in 2009. Other therapies such as activated protein C⁴⁹ and pentoxfylline^31,32 (currently being evaluated in a phase II trial as an adjunctive treatment for necrotizing enterocolitis (NEC) in premature infants [clinicaltrials.gov trial #NCT00271336]) have shown some promise, although future evaluation is warranted as their mechanisms of action are incompletely characterized and dangerous side effects (increased risk of bleeding) remain a concern with activated protein C⁵⁰. Other potential immunomodulatory targets include reduction of oxidative stress and amino acid supplementation during periods of critical illness. Melatonin, which reduced oxidative stress in neonatal animal models⁵¹ and free radical production in human neonates⁵², was shown to reduce markers of inflammation and sepsis mortality in a very small study⁵². Further investigation may be warranted in larger randomized controlled clinical trials to better assess the potential benefit of melatonin. When glutamine, an amino acid important for repair and growth of rapidly dividing tissues, was supplemented to critically ill adults, reduced rates of infection were noted^53–55. However, glutamine supplementation did not reduce nosocomial infections in preterm infants^30,56 or morbidity and mortality⁵⁷. As a ‘therapy’ with known immunologic benefit^6,58,59, human milk feeding has been associated with reduced infection in both term and preterm infants^6,60–63, and should be actively pursued.

Table 3. Attempts at Neonatal Immunomodulation and Impact on Sepsis Survival.

Numbers in parentheses represent references at the end of the text.

Immunomodulation Strategy	Proposed Benefit	Clinical Impact
Granulocyte transfusion	Increase functional granulocytes (PMNs)	No change in sepsis mortality (21)
GM-CSF, G-CSF	Stimulate proliferation and maturation of myeloid precursors in the bone marrow and modulate functions of mature PMNs at the site of infection	No change in sepsis mortality (34–38)
IVIG (monoclonal)	Increase bacteria-specific serum antibody titers	No change in sepsis mortality or efficacy of sepsis prevention (23,24,40,42)
IVIG (polyclonal)	Bind to cell surface receptors, provide opsonic activity, activate complement, promote antibody dependent cytotoxicity	Small (~3%) decrease in sepsis incidence, meta-analysis showed reduction in sepsis mortality with borderline statistical significance, INIS trial in progress (25,26,39,41,43–48)
Probiotics	Increase barrier to translocation of bacteria and bacterial products across mucosa, competitive exclusion of potential pathogens, modification of host response to microbial products	Limited data, reduction in NEC and NEC/sepsis mortality (27)
Breast milk	Multiple mechanisms: contains cellular defenses, IgA, APPs, probiotics, gut growth factors	Clear proven reduction in neonatal infectious mortality (28,58–63)
Activated protein C	Mechanism unclear	Limited data, no change in sepsis mortality, safety concerns (29,49–50)
Glutamine	Enhance gut integrity and immune function, and decrease bacterial translocation	No change in sepsis mortality (30,53,56,57)
Pentoxifylline	Reduce TNF-α and IL-6 synthesis	Limited data, reduction in sepsis mortality (31,32)
Anti-endotoxin antibodies	Decrease detrimental effects of endotoxin	Single small study, no change in mortality (33)

Role of probiotics in immunomodulation

As the largest immune organ, the intestine serves as the major nexus of interaction with the external environment. The innate immune system of the intestine serves as a selective barrier to the post-natal presence of microbes and food antigens in order to prevent overwhelming systemic infection. At birth, the sterile intestine is rapidly colonized with microorganisms from maternal and environmental sources. Commensal microbes participate in health of the individual by playing roles in nutrition and in gut development⁶⁴. Bacterial colonization of preterm infants differs from that of healthy full-term infants because of the methods of neonatal care (widespread prophylaxis with antibiotics, parenteral nutrition, and feeding in incubators) that may delay or impair colonization. The effect of antibiotics on the developing GI tract is incompletely understood, but recent studies in neonatal rodents demonstrate that the use of the ampicillin-like antibiotic, clomoxyl, alters development of intestinal gene expression⁶⁵. The short and long-term effects of these antibiotic-driven alterations in the GI microbiota (>500 species of microbes are known to reside in the human gastrointestinal tract^64,66) are the subject of intense research. The interaction between the mucosal immune system and the microbiota may be critical and the absence of this relationship may play a role in the development of autoimmune pathology and allergy. Probiotics have been successfully used as immunomodulators in the prevention and treatment of allergies in children^67,68, suggesting that modification of the relationship between the intestine and commensal organisms helps shape the immunological network of the host and therefore could serve as a target for immunomodulation in premature neonates.

Recent studies in mice demonstrating an increased susceptibility to hemorrhagic colitis after eradication of intestinal bacteria suggest that the epithelium and resident immune cells do not simply tolerate commensal organisms but are dependent on them for protection⁶⁹. Commensal bacteria express molecules such as lipopolysaccharide (LPS; Gram-negative bacteria), lipoteichoic acid (LTA; Gram-positive bacteria), and bacterial lipoproteins (BLPs, both Gram-negative and Gram-positive bacteria), that engage intestinal epithelial TLRs. The resultant tonic TLR signaling, enhances the ability of the epithelial surface to withstand chemical or inflammatory mediator-induced injury while also priming the surface for enhanced repair responses^69,70 (Fig. 1). This is due in part to selective coupling of TLR pathways in intestinal epithelial cells to production of anti-inflammatory and repair-enhancing cytokines⁷¹. The importance of TLR expression and stimulation in the development of NEC was highlighted in a recent report in neonatal mice indicating that over-expression of TLR4, following hypoxia or remote infection, was associated with a decreased ability of the intestine to withstand damage as well as an impairment in intestinal repair mechanisms^72,73. Thus, either the disruption of gut TLR signaling or the removal of TLR agonists compromises the ability of the intestinal surface to protect and repair itself in the face of an inflammatory or infectious insult^69,70,74 increasing the risk of bacterial translocation and the development of endotoxemia, sepsis, and/or NEC.

Figure 1. Interaction of Microbiota and Neonatal Intestine.

The interaction between intestinal bacteria and intestinal epithelium is necessary for homeostasis and normal function of repair mechanisms. Disruption of this interaction, through the use of antibiotics or via stress to the organism (i.e. remote infection or hypoxia), results in loss of homeostasis and degradation of the intestinal boundaries with subsequent microbial translocation.

Preterm human infants randomly assigned to receive a daily feeding supplement of a probiotic mixture compared to non-probiotic administered control infants had a relative risk reduction in mortality, incidence of NEC, and late onset sepsis^27,75. However, a large multicenter trial conducted in 12 Italian neonatal intensive care units on 565 patients did not elicit a statistically significant beneficial effect of probiotic (Lactobacillus GG) on NEC or sepsis⁷⁶. In a recent meta-analysis, probiotic treatment was shown to reduce the incidence of higher Bell stage NEC (stage 2 and higher) and mortality in preterm infants less than 1500 grams⁷⁷. Whether the differences in outcomes in these studies are associated with the use of different probiotic preparations, a different baseline incidence of NEC and sepsis in the different NICUs, or other factors such as breast milk feeding remain speculative. Previous studies suggest that human breast milk contains beneficial microbes that are independent of those found in the areolar skin⁷⁸. The role of human milk, itself replete with innate immune effectors including APP and TLR modulators^79,80, versus formula feedings in the probiotic studies mentioned above was not critically examined but is clearly worthy of future study. Additional studies are also necessary to clarify whether types of probiotics (live or inactivated bacteria or their corresponding TLR agonists) and their dosage have a differential impact on NEC prevention and/or sepsis survival, as well as on long-term outcome measures including safety.

Where do we go from here? In vitro studies light the way…

Many in vitro studies of neonatal leukocytes have identified functional deficiencies, pointing to potential future immunomodulatory therapies that may enhance specific facets of neonatal immune function, such as cytokine and APP production as well as phagocytosis. To facilitate the translation of this in vitro data to the bedside, preclinical models in animals have been employed. For example, studies of sepsis in adult animals have guided biopharmaceutical development of novel sepsis therapies including probiotics⁸¹, caspase inhibitors¹⁸, glucocorticoid-induced TNF-α receptor stimulation¹⁹, TLR modulation^82–84, as well as glucocorticoids, anti-endotoxin antibodies, IL-1 receptor antagonists, anti-TNF-α antibodies, recombinant bactericidal/permeability-increasing protein (BPI)⁸⁵, and others⁸⁶. However, much of the descriptive and mechanistic work characterizing the immune response in sepsis has been done using adult animal models that may not necessarily apply to neonates given the distinct function of neonatal immune system¹².

Innate immune system enhancement via TLR stimulation

Following the description of TLRs in humans⁸⁷, the characterization of their respective effects has grown exponentially. Widely expressed by invertebrate and vertebrate animals, TLRs detect pathogen-associated molecular patterns (PAMPs), triggering host cell cytokine and antimicrobial effector responses^88,89 (Fig. 2). TLR-triggered responses are differentially regulated at the genetic level and constitute an adaptive component of the innate immune system. For example, upon restimulation of leukocytes with TLR agonists, potentially harmful excessive inflammatory cytokine production is reduced while antimicrobial genes remain active⁹⁰. The contribution of TLR activation, signaling, and expression to human disease is currently an area of intensive study across many medical disciplines⁹¹.

Figure 2. Toll-Like Receptors (TLRs) and Their Respective Agonists.

Schematic representation of toll-like receptors, both cell surface and endocytic, as well as their respective ligands. Stimulation of these receptors causes enhanced immune function through increased cellular cytokine production and improvement of antimicrobial effector mechanisms such as cellular phagocytosis and antimicrobial protein and peptide (APP) production. LTA-lipotechoic acid. PG-peptidoglycan. LPS-lipopolysaccharide.

In the context of well-described in vitro deficits in neonatal immunity, the TLRs prominent role in innate immunity sparked studies to assess whether TLR activation improves neonatal cellular responses^92–98. Major differences exist between newborns and adults in TLR-mediated monocyte activation, but that these differences are not uniform across all TLRs⁹⁹. Although neonatal APCs demonstrate impaired production of T_H1-polarizing cytokines (eg, TNF-α and IL-12) and impaired up-regulation of co-stimulatory molecules in responses to agonists of TLRs 1–7, responses to TLR8 agonists are remarkably preserved^99–101. Impaired TLR signaling also contributes to reduced neonatal leukocyte responses to bacterial cell wall products, potentially contributing to neonatal risk of infection⁹⁷.

Impairment in the ability of human neonatal blood monocytes to produce TNF-α in response to TLR agonists is due to the inhibitory effects of plasma adenosine, which binds its specific adenosine receptor thereby increasing intracellular concentrations of the second messenger cyclic adenosine monophosphate (cAMP)^102,103. Neonatal blood plasma contains relatively high concentrations of adenosine and neonatal cord blood mononuclear cells demonstrate increased sensitivity to the cAMP-mediated inhibitory effects of adenosine¹⁰⁴. cAMP inhibits production of multiple T_H1 polarizing cytokines while preserving production of counter-regulatory cytokines such as IL-10. Thus, the adenosine system thus plays important roles in limiting TLR2-mediated blood monocyte production of TNF-α and other T_H1 polarizing cytokines at birth, during exposure of the neonate to microbial flora during the birth process¹². TLR8 agonists, including single stranded RNAs and synthetic imidazoquinoline compounds, represent an important exception to he general pattern of impaired human neonatal monocyte responses to TLR agonists^101,105.

Recent in vivo studies have explored whether priming of an innate immune response by pre-treatment of neonatal mice with TLR agonists improves outcomes of infection. Priming using TLR9 agonist (CpG) was protective against L. monocytogenes or neurotropic Tacaribe arenavirus (a powerfully lethal murine-specific virus) in single-pathogen challenge murine neonatal sepsis models^106,107. Using a neonatal murine model of polymicrobial sepsis¹⁰⁸, we showed that a single small dose prior to sepsis with TLR7/8 agonist (the imidazoquinoline resiquimod), or TLR4 agonist (LPS), was able to reduce polymicrobial sepsis mortality by 25 and 40% respectively over sham-treated septic animals (In press-Wynn et. al. Blood). Importantly, we found TLR agonist priming was not uniformly beneficial, as TLR3 agonist pretreated mice had no advantage over sham-treated septic animals. TLR7/8 and TLR4 agonist pretreatment enhanced but shortened the systemic inflammatory response seen with sepsis, decreased bacteremia, and improved survival, but were associated with associated with distinct enhancement of innate immune function including increased PMN recruitment, and reactive oxygen species production [TLR4] or phagocytic function [TLR7/8] (Fig. 3). In contrast to results in adult mice, in which the absence of the adaptive immune system was associated with significantly higher mortality (~60% higher than wild-type), there was no difference in neonatal sepsis mortality when compared to the wild-type mice (In press-Wynn et. al. Blood), emphasizing the dependence of the neonate on innate immunity for sepsis survival. Lastly, we showed that the survival benefit conferred by pretreatment with TLR agonists were also evident in neonatal mice lacking an adaptive immune system, confirming that the benefits of TLR agonist pretreatment are through stimulation of innate immunity alone. Thus, in marked contrast to adults, in whom absence or dysfunction of the adaptive immune system negatively affects survival following polymicrobial sepsis, neonates demonstrate a dependence on innate immune function that is improved through select TLR stimulation resulting in increased sepsis survival with minimal or no contribution of the adaptive immune system.

Figure 3. Effect of TLR agonist Pretreatment on Sepsis Survival, Cytokine Production, Bacteremia, and Innate Cellular Function in Neonatal Mice.

Systemic administration of select TLR agonists prior to the initiation of experimental polymicrobial sepsis significantly reduces murine neonatal mortality and is associated with a more robust but shortened inflammatory response and unique improvements in cell function (increased phagocytosis and reactive oxygen species production) with associated decreases in bacteremia as compared to saline pretreated neonatal mice.

Potential for Prophylactic Immunomodulation

Efforts aimed at improving neonatal sepsis survival by using adjuvants must take into consideration the unique aspects of neonatal immunity and sepsis in order to develop age-appropriate therapies. Specific TLR stimulation appears to effectively address some of the known deficits unique to neonatal immunity, including the long-appreciated impairment in phagocytic function during stress conditions^12,109,110. The efficacy of TLR agonist pretreatment in improving survival in preclinical models of severe infection is thus of great interest. Such a prophylactic approach is particularly appropriate for low-birth weight extremely premature neonates, a group at high risk for infection¹¹¹ due to particularly impaired innate immune responses (eg, skin immaturity, particularly low complement and APP levels, and deficiency of passive immunoglobulins) as well as exposure to multiple invasive procedures, including intubation, central venous line placement, delayed feeding, and monitoring devices. Premature infants suffer the highest sepsis mortality of any population⁵ and early onset lethal sepsis (during the first week of life) is predominantly due to Gram-negative bacteria¹¹¹. If a prophylactic protective immune response could be generated that would persist through this crucial time, it is likely that the mortality within this population could be dramatically reduced. As a result, potential prophylactic approaches directed at priming the neonatal immune system before a potential insult are of substantial interest. Priming innate immunity might also serve other immunocompromised medical populations such as those who have received chemotherapy or those who are pharmacologically immunosuppressed.

Potential Adverse Effects of Immunomodulation

Immunomodulation is not without theoretical risk of unwanted side effects. Improvements in innate cellular function might come at the expense of increased autoimmunity or other inflammatory damage. Because neonates with antenatal conditions which produce excessive inflammation (such as chorioamnionitis) are known to have associated pathology such as periventricular leukomalacia or chronic lung disease¹¹², the notion of augmenting the immune system response must be tempered by concerns for potential untoward effects. The recent discovery that effects of TLR stimulation, including cytokine production and antimicrobial effector mechanisms, are differentially regulated suggests that selective immunomodulatory effects can be elicited through specific pharmacologic targeting⁹⁰. TLR agonists, such as Bacillus Calmette-Guerin (a TLR2 and 4 agonist¹¹³), are already given to neonates at birth to reduce tuberculosis. More studies examining the effects of TLR stimulation in vivo are necessary in neonatal animal models, including non-human primates, whose TLR pathways more closely conform to those of humans¹¹⁴, with emphasis on determination of pharmacokinetics, pharmacodynamics, and potential side effects.

Summary

Modulation of innate immune responses provides fresh opportunities for improving sepsis survival in neonates. Targeting high risk populations for selective immunoprophylaxis to enhance host defense against bacterial infection is a promising approach whose realization will require further progress regarding mechanisms of action, pharmacologic properties, and safety.