Chapter 33: Enterobacteriaceae

BACTERIOLOGY

The Enterobacteriaceae are among the largest bacteria, measuring 2 to 4 μm in length with parallel sides and rounded ends. Forms range from large coccobacilli to elongated, filamentous rods. The organisms do not form spores or demonstrate acid-fastness.

The cell wall, cell membrane, and internal structures are morphologically similar for all Enterobacteriaceae, and follow the cell plan described in Chapter 21 for gram-negative bacteria. Components of the cell wall and surface, which are antigenic, have been extensively studied in some genera and form the basis of systems dividing species into serotypes. The outer membrane lipopolysaccharide (LPS) is called the O antigen. Its antigenic specificity is determined by variation in the sugars that form the long terminal polysaccharide side chains linked to the core polysaccharide and lipid A. Cell surface polysaccharides may form a well-defined capsule or an amorphous slime layer and are termed the K antigen (from the Danish kapsel, capsule). Motile strains have protein peritrichous flagella, which extend well beyond the cell wall and are called the H antigen. Many Enterobacteriaceae have surface pili (fimbriae), which are antigenic proteins, but not part of formal typing systems.

Enterobacteriaceae grow readily on simple media, often with only a single carbon energy source. Growth is rapid under both aerobic and anaerobic conditions, producing 2 to 5 mm colonies on agar media and diffuse turbidity in broth after 12 to 18 hours of incubation. All Enterobacteriaceae ferment glucose, reduce nitrates to nitrites, and are oxidase-negative.

CLASSIFICATION

Genus and species designations are based on phenotypic characteristics, such as patterns of carbohydrate fermentation, and amino acid breakdown. The O, K, and H antigens are used to further divide some species into multiple serotypes. These types are expressed with letter and number of the specific antigen, such as E coli O157:H7, the cause of numerous foodborne outbreaks. These antigenic designations have been established only for the most important species and are limited to known antigenic structures. For example, many species lack capsules and/or flagella. In recent years, DNA and rRNA homology comparisons have been used to validate these relationships and establish new ones. The genera containing the species most virulent for humans are Escherichia, Shigella, Salmonella, Klebsiella, and Yersinia. Other less common but medically important genera are Enterobacter, Serratia, Proteus, Morganella, and Providencia. Multiple locus sequence typing (MLST) schemes have been developed in order to facilitate identification of and communication about important clones within these key pathogenic species.

TOXINS

In addition to the LPS endotoxin common to all gram-negative bacteria, some Enterobacteriaceae also produce protein exotoxins, which act on host cells by damaging membranes, inhibiting protein synthesis, or altering metabolic pathways. The end result of these actions may be cell death (cytotoxin) or a physiologic alteration, the net effect of which depends on the function of the affected cell. For example, enterotoxins act on intestinal enterocytes, causing the net secretion of water and electrolytes into the gut to produce diarrhea. Although these toxins are most strongly associated with E coli, Shigella, and Yersinia, others with the same or very similar actions have now been discovered in other species. Toxins found in another species may differ slightly in protein structure and genetic regulation, but still have the same biologic action on host cells. Details of these toxins are discussed later in this chapter in relation to their prototype species.

DISEASES CAUSED BY ENTEROBACTERIACEAE

EPIDEMIOLOGY

Most Enterobacteriaceae are primarily colonizers of the lower gastrointestinal tract of humans and animals. Many species survive readily in nature and live freely anywhere that water and minimal energy sources are available. In humans, they are the major facultative components of the colonic bacterial flora, but are also found in the female genital tract and as transient colonizers of the skin. Enterobacteriaceae are scant in the respiratory tract of healthy individuals; however, their numbers may increase in hospitalized patients with chronic debilitating diseases. Escherichia coli is the most common species of Enterobacteriaceae found among the indigenous microbiota, followed by Klebsiella, Proteus, and Enterobacter species. Salmonella and Shigella species are not considered members of the microbiota, although carrier states can exist. Shigella and Salmonella serovar Typhi are strict human pathogens with no animal reservoir. An overview of these infections is illustrated in Figure 33–1.

PATHOGENESIS

Opportunistic Infections

Enterobacteriaceae are often poised to take advantage of their presence in the environment and human microbiota to produce disease whenever they gain access to normally sterile body sites. Surface structures such as pili have been proven to aid this process for some species, and likely do so for many others as well. Once in deeper tissues, bacteria may persist and cause injury using strategies that, with the exception of LPS endotoxin and production of exotoxins or capsules, are still not well understood. The prototype opportunistic infection is the UTI, in which Enterobacteriaceae gain access to the urinary bladder due to minor trauma or instrumentation. Strains able to adhere to uroepithelial cells can persist and multiply in the nutrient-rich urine, sometimes ascending by way of the ureters to the renal pelvis and kidney, producing pyelonephritis. Likewise, mucosal or skin trauma can allow access to soft tissues and aspiration can provide access to the lung when the relevant sites are colonized with Enterobacteriaceae.

Intestinal Infections

Salmonella, Shigella, Yersinia enterocolitica, and certain strains of E coli are able to produce disease in the intestinal tract. These intestinal pathogens have invasive properties or virulence factors such as cytotoxins and enterotoxins, which correlate with the type of diarrhea they produce. In general, the invasive and cytotoxic strains produce an inflammatory diarrhea called dysentery with white blood cells (WBCs) and/or blood in the stool. The enterotoxin-producing strains cause a watery diarrhea in which fluid loss is the primary pathophysiologic feature. For a few species, the intestinal tract may be the original portal of entry, but the disease ultimately becomes systemic as a result of spread of bacteria to multiple organs. Enteric (typhoid) fever caused by Salmonella enterica ser. Typhi is the prototype of this form of infection.

Regulation of Virulence

In addition to adherence pili, LPS, and exotoxins, the Enterobacteriaceae produce a myriad of other virulence factors to cause disease. Many of them are deployed in a complex and sequential fashion in response to environmental signals (temperature, iron, calcium) or as yet unknown factors. Some members have injection (type III or IV) secretion systems that target human cells by delivering a syringe-like injection of multiple virulence factors into the cytoplasm of host cells.

The genes for these factors, located in the chromosome, plasmids, or both, are controlled by interactive regulators that seem to produce each virulence factor exactly when it is needed. The genes themselves are often organized into clusters, which include the genes for the effector molecules as well as their regulatory proteins. This is particularly true for complex characteristics such as invasiveness, which involve multiple sequential steps. Some of these gene clusters reside within pathogenicity islands (PAIs) acquired en bloc from another bacterium in the genetically distant past. In particular, PAIs are associated with injection secretion systems, where they contain the structural genes for the injection apparatus, as well as the virulence factors injected.

IMMUNITY

Little is understood about immunity to the broad range of opportunistic infections caused by Enterobacteriaceae. Antibody directed against an LPS core antigen has been shown to provide a degree of protection against gram-negative endotoxemia, but the diversity of antigens and virulence factors among the Enterobacteriaceae is too great to expect broad immunity in any given host. Immunity to intestinal infection is generally short-lived, and will be discussed where relevant to specific intestinal pathogens.

ENTEROBACTERIACEAE: CLINICAL ASPECTS

MANIFESTATIONS

The Enterobacteriaceae produce the widest variety of infections of any group of microbial agents, including two of the most common infectious states, UTI and acute diarrhea. Urinary tract infections are manifested by dysuria and urinary frequency when infection is limited to the bladder, with the addition of fever and flank pain when the infection spreads to the kidney. Enterobacteriaceae are by far the most common cause of UTIs, and the most common species involved is E coli.

DIAGNOSIS

Culture is the primary method of diagnosis; all Enterobacteriaceae are readily isolated on routine media under almost any incubation conditions. Special indicator media such as MacConkey agar are commonly used in primary isolation to speed separation of the many species. For example, the common pathogens E coli and Klebsiella typically ferment lactose rapidly, producing acid (pink) colonies on MacConkey agar, whereas the intestinal pathogens Salmonella and Shigella do not. Separation of the intestinal pathogens from all the other Enterobacteriaceae in stool requires highly selective media designed solely for this purpose. (These are discussed as they relate to individual pathogens.) Improved understanding of the genetic and molecular basis for virulence has led to the development of direct nucleic acid and immunodiagnostic techniques for direct detection of toxin, adhesin, and invasin proteins or their genes in clinical material (eg, stool). Once too expensive for use in clinical laboratories, these methods are beginning to emerge as primary diagnostic tools.

TREATMENT

Antimicrobial therapy is crucial to the outcome of infections with members of the Enterobacteriaceae. Unfortunately, combinations of chromosomal and plasmid-determined resistance render them the most variable of all bacteria in susceptibility to antimicrobial agents. They are intrinsically resistant to penicillin G, erythromycin, and clindamycin, but may be susceptible to the extended spectrum β-lactams, carbapenems, aminoglycosides, tetracycline, chloramphenicol, sulfonamides, quinolones, nitrofurantoin, and the polypeptide antibiotics. The emergence and spread of multidrug-resistance plasmids featuring hydrolytic enzymes such as extended-spectrum cephalosporinases or carbapenemases have fundamentally altered the empiric treatment of gram-negative infections. Because the probability of resistance varies among genera and in different epidemiologic settings, the susceptibility of any individual strain must be determined by antimicrobial susceptibility tests. Typical patterns of resistance for some of the more common Enterobacteriaceae appear in Appendix 23–1.

BACTERIOLOGY

Most strains of E coli ferment lactose rapidly and produce indole. These and other biochemical reactions are sufficient to separate it from the other species. There are over 150 distinct O antigens and a large number of K and H antigens, all of which are designated by number. The antigenic formula for serotypes is described by linking the letter (O, K, or H) and the assigned number of the antigen(s) present (eg, O111:K76:H7). Alternatively, the emergence of multidrug-resistant clones in the sequencing era has led to the use of newer descriptive terminology (eg, E coli Sequence Type 131, Klebsiella pneumoniae Sequence Type 258).

PILI

Pili play a role in virulence as mediators of attachment to human epithelial surfaces. They show marked tropism for different epithelial cell types, which is determined by the availability of their specific receptor on the host cell surface. Most E coli express type 1 or common pili. Type 1 pili bind to the D-mannose residues commonly present on epithelial cell surfaces and thus mediate binding to a wide variety of cell types. More specialized pili are found in select clones of E coli. P pili bind to digalactoside (Gal–Gal) moieties on kidney cells and erythrocytes of the P blood group. Pili that mediate binding to enterocytes are found among the diarrhea-causing E coli and are specific to the pathogenic type, as shown in Figure 33–2 and listed in Table 33–1. Escherichia coli also causes diarrhea in animals, and different sets of pili exist with host-specific tropism for their enterocytes. The receptor(s) for the enteric pili are not known in detail, but include glycolipids and glycoproteins on the enterocyte surface.

The genetics of pilin expression are complex. The genes are organized into multicistronic clusters that encode structural pilin subunits and regulatory functions. Pili of different types may coexist on the same bacterium, and their expression may vary under different environmental conditions. Type 1 pilin expression can be turned “on” or “off” by inversion of a chromosomal DNA sequence containing the promoter responsible for initiating transcription of the pilin gene. Other genes control the orientation of this switch.

TOXINS

As a single species, E coli can produce every kind of protein exotoxin found among the Enterobacteriaceae. These include a pore-forming cytotoxin, inhibitors of protein synthesis, and a number of toxins that alter messenger pathways in host cells. The α-hemolysin is a pore-forming cytotoxin that inserts into the plasma membrane of a wide range of host cells in a manner similar to streptolysin O (Chapter 25) and Staphylococcus aureus α-toxin (Chapter 24). The toxin causes leakage of cytoplasmic contents and eventually cell death. The more recently discovered cytotoxic necrotizing factor (CNF) is often produced in concert with α-hemolysin. CNF is an A–B toxin that disrupts G proteins regulating signaling pathways in the cell cytoplasm with multiple effects including cytoskeleton rearrangement and apoptosis.

Shiga toxin (Stx) is named for the microbiologist who discovered Shigella dysenteriae, and this toxin was once believed to be limited to that species. It is now recognized to exist in at least two molecular forms released upon bacterial lysis by multiple E coli and Shigella strains. In the years after the discovery of this toxin, the term Shiga toxin was reserved for the original toxin, while others were called Shiga-like. In this book, the term Stx is used for all molecular variants that have the same mode of action regardless of the species under consideration. Stx is an A–B type toxin. The B unit directs binding to a specific glycolipid receptor (Gb₃) present on eukaryotic cells and leads to internalization by an endocytotic vacuole. Inside the cell, the A subunit crosses the vacuolar membrane in the trans-Golgi network, exits to the cytoplasm, and enzymatically modifies the ribosome site (28S-RNA of 60S subunit) where amino acyl tRNA binds. This alteration blocks protein synthesis, leading to cell death (Figure 33–3). This action is very similar to the plant toxin ricin.

Labile toxin (LT) is also an A–B toxin. Its name relates to the physical property of heat lability, which was important in its discovery, and contrasts with the heat-stable toxin (ST) also produced by E coli. The B subunit binds to the cell membrane, and the A subunit catalyzes the ADP-ribosylation of a regulatory G protein located in the membrane of the intestinal epithelial cell. This inactivation of part of the G protein complex causes permanent activation of the membrane-associated adenylate cyclase system and a cascade of events, the net effect of which depends on the biologic function of the stimulated cell. If the cell is an enterocyte, the result is the stimulation of chloride secretion out of the cell and the blockage of NaCl absorption. The net effect is the secretion of water and electrolytes into the bowel lumen. The structure and action of LT are nearly identical with that already described for cholera toxin (CT), but LT is less potent than CT.

Stable toxin is a small peptide that binds to a glycoprotein receptor, resulting in the activation of a membrane-bound guanylate cyclase. The subsequent increase in cyclic GMP concentration causes an LT-like net secretion of fluid and electrolytes into the bowel lumen.

E COLI OPPORTUNISTIC INFECTIONS

URINARY TRACT INFECTION

Epidemiology

Escherichia coli accounts for more than 90% of the more than 7 million cases of cystitis and 250 000 of pyelonephritis estimated to occur in otherwise healthy individuals every year in the United States. Urinary tract infections are much more common in women, 40% of whom have an episode in their lifetime, usually when they are sexually active. The reservoir for these infections is the patient’s own intestinal E coli microbiota, which contaminate the perineal and urethral area. In individuals with urinary tract obstruction or instrumentation, exogenous sources assume greater importance.

Pathogenesis

Relatively minor trauma or mechanical disruptions can allow bacteria colonizing the periurethral area brief access to the urinary bladder. These bacteria originally derived from the fecal microbiota are frequently present in the bladder of women immediately after sexual intercourse. In most instances, they are purged by the flushing action of voiding, but may persist to cause a UTI, depending on host and bacterial factors. Although most UTIs are in otherwise healthy women, host situations that violate bladder integrity (urinary catheters) or that obstruct urine outflow (in men, enlarged prostate) may allow the bacteria more time to attach, multiply, and cause injury. Here, bacterial virulence factors are important, and E coli is the prototype UTI pathogen. Fewer than 10 E coli clones (whether characterized as serotypes or sequence types) account for the majority of UTI cases, and these UTI clones are not the most common ones in the microbiota. These E coli with enhanced potential to produce UTI are called uropathogenic E coli (UPEC).

The ability of UPEC to produce UTI begins with type 1 pili, which are the most important for both periurethral and bladder colonization. The tips of such pili attach to mannose moieties presented by membrane proteins (uroplakins) in the transitional epithelium of the bladder. Other pili such as P pili may add to the strength of this attachment; however, because their cognate Gal–Gal receptor is most abundant in the renal pelvis and kidney, P pili are more important for upper urinary tract disease. E coli possessing P pili are a minor percentage of the fecal flora (less than 20%), but the proportion of P⁺ strains progressively rises with the severity of UTI, up to 70% in pyelonephritis isolates. Motility driven by flagellar motors also plays a role both in access to the bladder and swimming up the ureter to the kidney. Obviously, adherence and motility are at cross-purposes, but UPEC are able to reciprocally regulate these features with on/off switching of pilin expression; but even with type 1 pili switched on, UPEC can alternate between swimming and adherent phases due to the catch-bond properties of the tip adhesin. Another feature is the ability of UPEC to invade superficial epithelial cells. The raft-like clusters formed by this maneuver are felt to aid persistence against the periodic flushing of the bladder. Once established, LPS and the production of other virulence factors such as α-hemolysin and CNF cause injury. Spread to the bloodstream leads to LPS-induced septic shock. The adherence aspects of UPEC are illustrated in Figure 33–4.

OTHER OPPORTUNISTIC INFECTIONS

Meningitis

Escherichia coli is one of the most common causes of neonatal meningitis, producing many features similar to group B streptococcal disease. The pathogenesis involves colonization of the infant with maternal E coli via ruptured amniotic membranes or during childbirth. Failure of protective maternal IgM antibodies to cross the placenta and the immunologic immaturity of newborns surely play a role. Fully 75% of cases of newborn meningitis are caused by strains possessing the sialic acid-containing K1 capsular polysaccharide that is structurally identical to the group B polysaccharide of Neisseria meningitidis, another cause of meningitis.

With the exception of UTIs, extraintestinal E coli infections are uncommon unless there is a significant breach in host defenses. Opportunistic infection may follow mechanical damage such as trauma or a ruptured intestinal diverticulum, or involve a generalized impairment of immune function. The virulence factors involved are likely the same as with UTI (eg, pili, α-hemolysin), but have been less specifically studied. Failure of local control of infection can lead to spread and eventually gram-negative septic shock. A significant proportion of blood isolates have the K1 surface polysaccharide. The particular diseases that result depend on the sites involved.

E COLI INTESTINAL INFECTIONS

Diarrheal illnesses continue to produce a tremendous mortality burden worldwide, particularly in children under 5 years of age, with the largest numbers of deaths occurring in sub-Saharan Africa and South Asia. Escherichia coli and Shigella (which are specialized E coli) are among the top causes of moderate to severe diarrhea among children in these areas. Diarrhea-causing E coli are classified according to their virulence properties as enterotoxigenic (ETEC), enteropathogenic (EPEC), enteroinvasive (EIEC), enterohemorrhagic (EHEC), or enteroaggregative (EAEC). Each group causes disease by a different mechanism, and the resulting syndromes usually differ clinically and epidemiologically. For example, ETEC and EIEC strains infect only humans. Food and water contaminated with human waste and person-to-person contact are the principal means of infection. A summary of the pathogenesis of infection, clinical syndromes, and epidemiology of infection for each enteropathogen is shown in Table 33–1.

ENTEROTOXIGENIC E COLI (ETEC)

Epidemiology

Enterotoxigenic E coli produce diarrhea in infants in developing countries, where they are a leading cause of morbidity and mortality during the first 2 years of life. ETEC is also the most important cause of traveler’s diarrhea in visitors to these countries. Repeated bouts of diarrhea caused by ETEC and other infectious agents are an important cause of growth retardation, malnutrition, and developmental delay in third-world countries where ETEC are endemic. ETEC disease is rare in industrialized nations, although recent outbreaks suggest that it may be underestimated.

Transmission is by consumption of food and water contaminated by infected or convalescent human carriers. Uncooked foods such as salads or marinated meats and vegetables are associated with the greatest risk. Direct person-to-person transmission is unusual, because the infecting dose is high. Animals are not involved in ETEC disease.

Pathogenesis

ETEC diarrhea is caused by strains of E coli that produce LT and/or ST enterotoxins in the proximal small intestine. ST seems to be more potent than LT, and strains that elaborate both cause the most severe illness. Adherence to surface microvilli, mediated by multiple variants of colonizing factor (CF) pili, is essential for the efficient delivery of toxin to the target enterocytes. The genes encoding the ST, LT, and the CF pili are borne on plasmids; a single plasmid can carry all three sets of genes. The bacteria remain on the epithelial surface, where the adenylate cyclase-stimulating action of the toxin(s) creates the flow of water and electrolytes from the enterocyte into the intestinal lumen. The mucosa becomes hyperemic but is not injured in the process. There is no invasion or inflammation.

Immunity

Although individuals may experience more than one episode of ETEC diarrhea, infections with ETEC do stimulate immunity. Travelers from industrialized nations have a much higher attack rate than adults living in the endemic area. This natural immunity is presumably mediated by secretory IgA specific for LT and CFs; the small ST peptides are nonimmunogenic. The disease is of very low incidence in breastfed infants, underscoring the protective effect of maternal antibody and the importance of transmission by contaminated food and water.

ENTEROPATHOGENIC E COLI (EPEC)

Epidemiology

Enteropathogenic E coli strains were first identified as the cause of explosive outbreaks of diarrhea in hospital nurseries in the United States and Great Britain during the 1950s. The link to E coli was established on epidemiologic grounds alone, using serotyping of stool isolates—no small task. The World Health Organization still recognizes a group of 12 EPEC serotypes. The disease seems to have disappeared in industrialized nations, although it may be underestimated because of the difficulty of diagnosis. In developing countries throughout the world, EPEC account for up to 20% of diarrheal illnesses in bottle-fed infants younger than 1 year of age. The reservoir is infant cases and adult carriers, with transmission by the fecal–oral route. Nursery outbreaks demonstrate the importance of spread by fomites, suggesting that the infecting dose for infants is low. Adult cases are felt to require a very high infecting dose (10⁸ to 10¹⁰ bacteria).

Pathogenesis

Enteropathogenic E coli initially attach to small intestine enterocytes using bundle-forming pili (Bfp) to form clustered microcolonies on the enterocyte cell surface. The lesion then progresses with localized degeneration of the brush border, loss of the microvilli, and changes in the cell morphology including the production of dramatic “pedestals” with the EPEC bacterium at their apex. The combination of these actions is called the attaching and effacing (A/E) lesion (Figure 33–5). The many steps involved in the formation of the A/E lesion are genetically controlled in a PAI, which includes the genes for the major EPEC attachment protein, intimin, and an injection (type III) secretion system. The secretion system injects over 30 E coli secretion proteins (Esps) into the host cell cytoplasm including—remarkably—the surface receptor (Tir) for intimin, which migrates to the surface after its injection. The other E coli secretion proteins perturb intracellular signal transduction pathways, one effect of which is the induction of modifications in enterocyte cytoskeleton proteins (actin, talin). The cytoskeleton accumulates beneath the attached bacteria to form the pedestals and complete the actin-rich A/E lesion (Figure 33–6). The Esps cause a host of other intracellular disruptions, including mitochondrial injury and induction of apoptosis. The link between these morphologic changes of the A/E and diarrhea is not known, but the injected Esps have also been shown to change electrolyte transport across the luminal membrane.

Immunity

In endemic areas, EPEC can be isolated often from the stool of asymptomatic adults, but unlike ETEC, these strains do not seem to cause traveler’s diarrhea in individuals new to the area. This observation obscures whether adults have acquired immunity or resistance based on physiologic factors.

ENTEROHEMORRHAGIC E COLI (EHEC)

Epidemiology

Enterohemorrhagic E coli disease and the associated hemolytic uremic syndrome (HUS) are the result of consumption of products from animals colonized with EHEC strains. It is also clear from secondary cases in families during outbreaks that person-to-person transmission also occurs. This disease occurs more in developed than developing countries.

EHEC was first recognized when outbreaks of HUS (hemolytic anemia, renal failure, and thrombocytopenia) were linked to a single E coli serotype, O157:H7. Since then, EHEC disease has emerged as an important cause of bloody diarrhea in industrialized nations and retained a remarkable but not exclusive relationship with the O157:H7 serotype. Regional and national outbreaks associated with ground beef, unpasteurized juices, and fresh vegetables often catch the attention of the public, the press, and the government, and at times bring renewed scrutiny to food safety policies and practices in public and private sectors.

The emergence of EHEC is related to its virulence, low infecting dose, common reservoir (cattle), and changes in the modern food processing industry that provide fresher meat (and bacteria) over wider distribution networks. The infecting dose, estimated to be as low as 100 organisms, is particularly important. This is a level at which food need not come directly from the infected animal, but only be contaminated by it. For example, large modern meat-processing plants can mix EHEC from colonized cattle at one ranch into beef from hundreds of other farms and quickly ship it all over the country. Therefore, the worst outbreaks have been seen in countries with the most advanced food production and distribution systems. If the organisms are ground into hamburger meat, an infecting dose of EHEC may remain even after cooking if the meat is left rare in the middle. Unpasteurized milk carries an obvious risk, but fruits and vegetables have also been the source for EHEC infection. In these instances, the EHEC from the manure of cattle grazing nearby has contaminated fruit in the field. The bacterial dose from a few “drop” apples (those picked up from the ground) included in a batch of cider has been enough to cause disease.

Pathogenesis

EHEC strains cause the A/E lesions previously described for EPEC, but also produce the Stx toxin. The EHEC pathotype, which first appeared in O157:H7 strains in 1982, is felt to have evolved by an EPEC acquiring the genes for Stx via prophage. Apparently the injection secretion system which creates the A/E pedestals also facilitates delivery of Stx to the enterocyte. Stx secretion is regulated through a quorum-sensing system which waits until there is a critical EHEC population to activate. The interaction of EHEC with enterocytes is much the same as that of EPEC, except that EHEC strains do not form localized microcolonies on the mucosa and have their own adhesive pili (long polar fimbriae [Lpf]), which mediate attachment in the colon rather than the small intestine. The outer membrane protein intimin mediates tight adherence, and the injection secretion system infuses the E coli secretion proteins, which cause alterations in the host cytoskeleton. The genes for these properties are also found in a PAI. The multiple extraintestinal features such as HUS are the result of circulating Stx.

The A/E features alone are sufficient to cause nonbloody diarrhea. On top of this, Stx production causes capillary thrombosis and inflammation of the colonic mucosa, leading to a hemorrhagic colitis. (The distinctive association between shiga toxin and bloody diarrhea has given rise to the term STEC, or shiga toxin-producing E coli, as an alternate pathotype descriptor for EHEC.) Although it has not been detected in the blood of human cases, Stx is presumed to be absorbed across the denuded intestinal mucosa. Circulating Stx binds to renal tissue, where its glycoprotein receptor (globotriaosylceramide, Gb3) is particularly abundant, causing glomerular swelling and the deposition of fibrin and platelets in the microvasculature. How Stx causes hemolysis is less clear; perhaps the erythrocytes are simply damaged as they attempt to traverse the occluded capillaries. Cases and outbreaks caused by Stx-producing E coli of other serotypes are common in many countries.

ENTEROINVASIVE E COLI (EIEC)

The biochemistry, genetics, and pathogenesis of enteroinvasive E coli strains are so close to those of Shigella that our understanding of EIEC disease is generally extrapolated from that genus—EIEC disease is essentially a mild version of shigellosis. Epidemiologically, EIEC infections are primarily seen in children younger than 5 years living in developing nations. The occasional documented outbreaks in industrialized nations are usually linked to contaminated food or water. There is a lower incidence of person-to-person transmission of EIEC, which correlates with the observation that the infecting dose is higher than it is for Shigella. Humans are the only known reservoir.

ENTEROAGGREGATIVE E COLI (EAEC)

Enteroaggregative E coli is associated with a protracted (more than 14 days) watery diarrhea that occasionally features blood and mucus. First recognized in infants and children in developing countries, EAEC is increasingly diagnosed in a variety of community settings. EAEC strains are identified by the “stacked-brick” pattern the bacteria make when adhering to cultured mammalian cells. The EAEC pili (aggregative adherence fimbriae [AAF]) mediate tight adherence to the intestinal mucosa, but the A/E lesions of the EPEC and EHEC are not present. The pathogenesis of diarrhea involves formation of a thick mucus–bacteria biofilm on the intestinal surface.

This view of EAEC was dramatically altered by a 2011 German outbreak of serotype O104:H4 initially thought to be caused by EHEC based on clinical features. There were a thousand cases of bloody diarrhea and 53 deaths due to HUS, but the rate of HUS development was twice that typical for EHEC disease. It turned out that the responsible strain had all the features of EAEC with the addition of Stx genes. There was no injection secretion system or A/E lesions. Apparently, the tight adherence of EAEC provided a particularly effective mechanism for delivery of Stx to the intestinal mucosa.

E COLI INFECTIONS: CLINICAL ASPECTS

MANIFESTATIONS

Opportunistic Infections

The most common symptoms of E coli UTI are dysuria and urinary frequency and do not differ significantly in character from those produced by the other less common gram-negative urinary pathogens. If the infection ascends the ureters to produce pyelonephritis, fever and flank pain are common and bacteremia may develop. Although E coli may have enhanced virulence in the production of pneumonia as well as soft tissue and other infections, no clinical features distinguish these cases from those caused by other members of the Enterobacteriaceae.

Intestinal Infections

Infections caused by all of the E coli virulence types usually begin with a mild watery diarrhea starting 2 to 4 days after ingestion of an infectious dose. In most instances, the duration of diarrhea is limited to a few days, with the exception of EAEC diarrhea, which can last for weeks. With ETEC and EPEC, the diarrhea remains watery, but with EIEC and EHEC, a dysenteric illness follows. Some EPEC cases may also become chronic. Enterohemorrhagic E coli disease begins like the others but often also includes vomiting. In 90% of cases, this is followed in 1 to 2 days by intense abdominal pain and bloody diarrhea, but fever is not prominent.

Some EHEC cases develop into a dysentery illness that is less severe than that seen in shigellosis. Colonoscopy reveals edema, hemorrhage, and pseudomembrane formation. Resolution usually takes place over a 3- to 10-day period, with few residual effects on the bowel mucosa.

Hemolytic uremic syndrome develops as a complication in 5% to 10% of cases of EHEC hemorrhagic colitis, primarily in children under 10 years of age. The disease begins with oliguria, edema, and pallor, progressing to the triad of microangiopathic hemolytic anemia, thrombocytopenia, and renal failure. The systemic effects are often life-threatening, requiring transfusion and hemodialysis for survival. The mortality rate is 5%, and up to 30% of those who survive suffer sequelae such as renal impairment or hypertension.

DIAGNOSIS

Like the rest of the Enterobacteriaceae, E coli is readily isolated in culture. In UTIs, the bacteria typically reach high numbers (greater than 10⁵/mL), which makes them readily detectable by Gram stain even in an unspun urine specimen (Figure 33–7). For the diagnosis of intestinal disease, separating the virulent E coli from those universally found in stool presents a special problem. A myriad of immunoassay and nucleic acid amplification methods have been described that are able to detect the toxins (LT, ST, Stx) or genes associated with virulence. These methods work but their clinical use is hampered by limited positive predictive value (healthy persons may also have positive test results with these methods) and high cost, especially in developing countries where ETEC, EIEC, EPEC, and EAEC are prevalent. A screening test for EHEC takes advantage of the observation that the O157:H7 serotype typically fails to ferment sorbitol. Incorporating sorbitol in place of lactose in MacConkey agar provides an indicator medium from which suspect (colorless) colonies can be selected and then confirmed with O157 antisera. This procedure has become routine in areas where EHEC is endemic but does not detect the non-O157 EHEC strains.

TREATMENT

Acute uncomplicated UTIs are often treated empirically. Because of widespread resistance to earlier agents like ampicillin, use of trimethoprim/sulfamethoxazole (TMP-SMX) and fluoroquinolones for this purpose rose steadily. In turn, use of these agents set the stage for the rise of international multidrug-resistant clones like E coli Sequence Type 131, which boasts chromosome-encoded fluoroquinolone resistance and plasmid-borne resistance to TMP-SMX, gentamicin and, not infrequently, extended-spectrum cephalosporins. Thus, in many clinical settings domestically and abroad, E coli resistance to these latter agents has now exceeded the 20% level used to indicate the suitability of antibiotics for empiric use. In cases of empiric treatment failure, selection of other antimicrobials must be guided by antimicrobial susceptibility testing of the patient’s isolate.

Because most E coli diarrheas are mild and self-limiting, treatment is usually not required. When it is, rehydration and supportive measures are the mainstays of therapy, regardless of the causative agent. In the case of EHEC with hemorrhagic colitis and HUS, heroic supportive measures such as hemodialysis or hemapheresis may be required. Treatment with TMP-SMX or fluoroquinolones reduces the duration of diarrhea in ETEC, EIEC, and EPEC infection. Because the risk of HUS may be increased by the use of antimicrobial agents, their use is contraindicated when EHEC is even suspected. Antimotility agents are not helpful and are contraindicated when EIEC or EHEC could be the etiologic agent.

PREVENTION

Traveler’s diarrhea is usually little more than an inconvenience. Because the infecting dose is high, the incidence of the disease can be greatly reduced by eating only cooked foods and peeled fruits and drinking hot or carbonated beverages. Avoiding uncertain water, ice, salads, and raw vegetables is a wise precaution when traveling in developing countries. High-priced hotel accommodations have no protective effect. Chemoprophylaxis against traveler’s diarrhea is not routinely recommended, and in fact may increase risk of intestinal colonization with multidrug-resistant Enterobacteriaceae encountered abroad. Short courses of TMP-SMX or ciprofloxacin (less than 2 weeks) have been recommended for those at high risk for disease resulting from such chronic conditions as achlorhydria, gastric resection, prolonged use of H₂ blockers or antacids, and underlying immunosuppressive diseases.

These public health measures apply equally to EHEC, but here prevention is more difficult because the infecting dose is so low. Cooking hamburgers all the way through is sensible, but no one is recommending abstinence from salads when at home. Recent US recommendations for the irradiation of meats and the extension of pasteurization requirements to fruit juices are designed largely to stem the spread of EHEC.

BACTERIOLOGY

Shigella species may be considered specialized E coli. Their antigenic makeup has been characterized in a manner similar to that of E coli, with the exception that they lack flagella and thus H antigens. All Shigella species are nonmotile. The genus is divided into four species, which are defined by biochemical reactions and specific O antigens organized into serogroups. The species are Shigella dysenteriae (serogroup A), Shigella flexneri (serogroup B), Shigella boydii (serogroup C), and Shigella sonnei (serogroup D). All but S sonnei are further subdivided into a total of 38 individual O antigen serotypes specified by numbers. Shigella is the prototype invasive bacterial pathogen. All species are able to invade and multiply inside a wide variety of epithelial cells, including their natural target, the enterocyte. Shigella dysenteriae type 1, the Shiga bacillus, is the most potent producer of Stx. Other Shigella species produce various molecular forms and quantities of Stx.

SHIGELLOSIS

EPIDEMIOLOGY

Shigellosis is a strictly human disease with no animal reservoirs. Worldwide, it is consistently one of the most common causes of infectious diarrhea with over 150 million cases and 600 000 deaths per year. As with almost all infectious diarrheas, the incidence is related to general levels of sanitation, but Shigella disease remains important in both developed and developing countries. This high prevalence despite lack of a nonhuman reservoir is primarily due to highly efficient transmission by the fecal–oral route. This spread by person-to-person contact is so effective because the infecting dose is extremely low, as few as 10 organisms in some studies. The secondary attack rates among family members are as high as 40%. Shigella is also spread by food or water contaminated by humans.

The incidence and spread of shigellosis is directly related to personal and community sanitary practices. In developed countries, it is largely a pediatric disease. In countries where the sanitary infrastructure is inadequate and in institutions plagued by crowding and poor hygienic conditions, the disease may be more widespread; wartime and natural disasters create similar circumstances. The most common species are S flexneri and S sonnei, with S dysenteriae largely limited to underdeveloped tropical areas. S dysenteriae type 1 produces the most severe disease, historically known as “bacillary dysentery.” This condition has slowed the march of many an army; it was the leading cause of death in the notorious Andersonville prison camp during the American Civil War.

PATHOGENESIS

Shigella, unlike Vibrio cholerae and most Salmonella species, is acid-resistant and survives passage through the stomach to reach the intestine. Once there, the fundamental pathogenic event is invasion and destruction of the human colonic mucosa. This triggers an intense acute inflammatory response with mucosal ulceration and abscess formation. The steps involved in this process describe one of the richest tales in bacterial pathogenesis (Figure 33–8). Most of the research has been done with S flexneri but there is no reason to believe it does not apply equally to the three other species and to EIEC.

Shigella initially cross the mucosal membrane by entering the lymphoid follicle-associated M cells of the intestine, which lack the highly organized brush borders of absorptive enterocytes. The Shigella adhere selectively to M cells, enter, and then transcytose through them into the underlying collection of macrophages. Inside macrophages, the organisms escape from the phagosome to the cytoplasm and activate programmed cell death (apoptosis) in the macrophage. Bacteria released from the dead macrophage contact the basolateral side of enterocytes and initiate a multistep invasion process mediated by a set of invasion plasmid antigens (IpaA–IpaD). On contact with the enterocyte, these proteins are injected by an injection (type III) secretion system and induce cytoskeleton reorganization, actin polymerization, and other changes particularly at the cell surface. Rather than create the A/E lesions of the EPEC and EHEC, this cytoskeleton modification process induces engulfment and internalization of Shigella into the host cell by endocytosis.

Shigella are highly adapted to the intracellular environment and make unique use of it to continue the infection. Although initially the bacteria are surrounded by a phagocytic vacuole, they quickly escape and enter the cytoplasmic compartment of the host cell. Almost immediately, they orient in parallel with the filaments of the actin cytoskeleton of the cell and initiate a process in which they control polymerization of the monomers that make up the actin fibrils. This process creates an actin “tail” at one end of the microbe, which appears to propel it through the cytoplasm like a comet. This exploitation of the cytoskeletal apparatus allows nonmotile Shigella to not only replicate in the cell but to move efficiently through it. Apparently, the cell’s microtubule network is an obstruction so the bacteria produce an enzyme that digests them. One microbiologist called this strategy “bushwhacking through a microtubule jungle.”

Eventually, the bacteria encounter the host cell membrane, much of which is adjacent to the neighboring enterocytes. At this point some Shigella rebound, but others push the membrane as much as 20 μm into the adjacent cell. This invasion of the neighboring enterocyte forms finger-like projections, which eventually pinch off, placing the bacterium within a new cell but surrounded by a double membrane. The organisms then lyse both membranes and are released into the cytoplasm, free to begin their relentless invasion anew.

The cell-by-cell extension of this process radially destroys enterocytes and creates focal ulcers in the mucosa, particularly in the colon. The ulcers add a hemorrhagic component and allow Shigella to reach the lamina propria, where they evoke an intense acute inflammatory response. Extension of the infection beyond the lamina (eg, to the bloodstream) is unusual in healthy individuals. The diarrhea created by this process is almost purely inflammatory, consisting of small-volume stools containing WBCs, RBCs, bacteria, and little else—this is classic dysentery.

Some Shigella also produce Stx, which is not essential for disease but does contribute to the severity of the illness. The original and most potent producer of Stx, S dysenteriae type 1, is the only Shigella with a significant mortality rate in previously healthy individuals. This is probably due to systemic effects of the toxin, which can be the same as previously described for the EHEC, including HUS. The role, if any, of Stx in enterocyte injury and diarrhea is uncertain.

All virulent Shigella and EIEC carry a very large plasmid that has several genes essential for the attachment and entry process, including the Ipa genes. The characteristics of Shigella entry and interaction with cellular elements are very similar to those observed with Listeria monocytogenes, which is gram positive and motile and prefers livestock to humans. Finding that such dissimilar bacteria use such similar tactics to infect their preferred host suggests commonality among the selective pressures on a microbe to become a “successful” enteric pathogen.

IMMUNITY

Episodes of Shigella infection produce modest immunologic protection against infection by homologous serotypes, but do not significantly cross-protect against other serotypes. Recently, large-scale epidemiologic studies have elucidated the predominant Shigella serogroups responsible for the major burden of disease, invigorating the prospects for design of a multivalent vaccine.

SHIGELLOSIS: CLINICAL ASPECTS

MANIFESTATIONS

Shigella organisms cause an acute inflammatory colitis and bloody diarrhea, which in the most characteristic state presents as a dysentery syndrome—a clinical triad consisting of cramps, painful straining to pass stools (tenesmus), and a frequent, small-volume, bloody, mucoid fecal discharge. However, most clinical shigellosis due to S sonnei is a watery diarrhea that is often indistinguishable from that of other bacterial or viral diarrheal illness. The disease usually begins with fever and systemic manifestations of malaise, anorexia, and sometimes myalgia. These nondescript symptoms are followed by the onset of watery diarrhea containing the large numbers of leukocytes detectable by light microscopy. The diarrhea may turn bloody with or without the other classic signs of dysentery. The manifestations may be more severe when S flexneri, the species that predominates in the developing world, is involved, and most severe with S dysenteriae type1 (Shiga bacillus). Although most cases of shigellosis resolve spontaneously after 2 to 5 days, the mortality rate in Shiga epidemics in Asia, Latin America, and Africa has been as high as 20%.

DIAGNOSIS

All Shigella species are readily isolated using selective media (eg, Hektoen enteric agar) that are part of the routine stool culture protocol in all clinical laboratories. These media contain chemical additives shown to inhibit facultative flora (eg, E coli, Klebsiella) with relatively little effect on Shigella (or Salmonella). They also contain indicator systems that use typical biochemical reactions to mark suspect Shigella colonies among the other flora. Isolates are identified with further biochemical tests. Slide agglutination tests using O group-specific antisera (A, B, C, D) confirm both the species and the Shigella genus.

TREATMENT

Several antimicrobial agents have proved effective in the treatment of shigellosis. Because the disease is usually self-limiting, the beneficial effect of treatment is in shortening the duration of the illness and the period of excretion of organisms. Ampicillin was once the treatment of choice, but resistance rates as high as 50% have caused a shift to other agents; in recent years, ciprofloxacin, ceftriaxone, and azithromycin have been used, depending on susceptibility testing. Antispasmodic agents may aggravate the condition and are contraindicated in shigellosis and other invasive diarrheas.

PREVENTION

Standard sanitation practices such as sewage disposal and water chlorination are important in preventing the spread of shigellosis. In certain circumstances, insect control may also be important, because flies can serve as passive vectors when open sewage is present. Good individual sanitary practices, such as handwashing and proper cooking of food, are highly protective. Parenteral vaccines have proved disappointing thus far. Current efforts encompass development of a wide range of constructs, including live attenuated, formalin-killed whole cell, glycoconjugate, subunit and novel antigen (such as Type III secretion systems and outer membrane protein) vaccines.

BACTERIOLOGY

More than any other genus, Salmonella has been a favorite of those who love to subdivide and apply names to biologic systems. At one time, there were over 2000 names for various members of this genus, often reflecting colorful aspects of place or circumstances of the original isolation (eg, S budapest, S seminole, S tamale, S oysterbeds). This rich nomenclature has now been streamlined to a single species, Salmonella enterica, with the previous species names relegated to the status of serovars. Adding to this robust picture are a large number of LPS O antigens, a few capsular K antigens, and flagellar H antigens that undergo phase variation (thus doubling the possible H antigenic states for each strain). As in Shigella, the specific O antigens are organized into serogroups (eg, A, B, …K, and so on), to which the two H and K (if present) antigen designations are appended to achieve the full antigenic formula. It is not difficult to understand why microbiologists—confronted by a salmonella with the antigenic formula O:group B [1,4,12] H:I;1,2—still prefer to call it Salmonella typhimurium. The proper name for this organism is Salmonella enterica serovar Typhimurium, but indulging in the convenience of elevating the serotype to species status is still common.

Another feature distinguishing Salmonella serotypes is their host range. Some are highly adapted to particular mammals or amphibians, and others infect a broad range of hosts. Of interest for medical microbiology are those strictly adapted to humans and those that infect humans and other animals. Salmonella enterica serovar Typhi is the prototype for the former, and S enterica serovar Typhimurium is the prototype for the latter. In the following discussions, Typhi is used for the strictly human species that produce enteric (typhoid) fever. Unless otherwise specified, S enterica is used for serotypes such as Typhimurium, which are able to infect animals or humans and typically cause gastroenteritis in the latter.

Salmonellae possess multiple types of pili, one of which is morphologically and functionally similar to E coli type 1 pili, and bind D-mannose receptors on various eukaryotic cell types. Most strains are motile through the action of their flagella. Salmonella Typhi has a surface polysaccharide called the Vi antigen, but capsules have not been important in the other Salmonella.

SALMONELLA GASTROENTERITIS (S ENTERICA)

EPIDEMIOLOGY

Salmonella enterica gastroenteritis is predominantly a disease of industrialized societies and improper food handling, which allows the transmission from the animal reservoir to humans. The infecting dose of S enterica infection varies widely with serotype (from 200 to 10⁶ bacteria), but is generally considerably higher than Shigella. This makes human-to-human transmission by direct contact unlikely, so these infections are transmitted by circumstances in which the bacteria increase their numbers by growth in contaminated foods before ingestion. Achlorhydric individuals or those taking antacids can be infected with smaller inocula.

Consistently, salmonellae are a leading cause of foodborne intestinal infection. The typical example of Salmonella “food poisoning” is the community picnic or bazaar, in which volunteers prepare poultry, salads, and other potential culture media to be eaten later in the day. Because the refrigerators are filled with beer and soda, the food is left out in covered pans. A near-physiologic incubation temperature is provided by the still-warm contents and the afternoon sun. This allows the organisms to enter logarithmic growth during the softball game. The bacteria usually produce no noticeable change in the food. One to two days after the feast, a significant proportion of the revelers develop abdominal pain, nausea, vomiting, and diarrhea lasting for 3 or 4 days. An investigation points to a particular food such as potato salad or turkey dressing, which is found to have a correlation with both attack rate and severity of illness.

Poultry products, including eggs (infected transovarially), are most often implicated as the vehicle of infection of Salmonella gastroenteritis. Food preparation practices that allow achievement of an infecting dose by growth of the bacteria in the food before ingestion are most commonly involved. The incidence in the United States is approximately double that of Shigella, with 40 000 to 50 000 reported cases per year, of an estimated total of 1.2 million illnesses annually. The number of cases varies seasonally, with peak incidence in summer and fall.

The highest rates of infection are in children under 5 years of age, persons aged 20 to 30, and those older than 70. If one household member becomes infected, the probability that another will become infected approaches 60%. Nearly one-third of all Salmonella epidemics occur in nursing homes, hospitals, mental health facilities, and other institutions. Increases in the popularity of raw milk have been associated with outbreaks of Salmonella (and Campylobacter) infection. Exotic pets such as turtles have also been the source of infection. Humans can also be the source of disease. Fully 5% of patients recovering from gastroenteritis still shed the organisms 20 weeks later. Chronic carriers who are food handlers are an important reservoir in the epidemiology of foodborne disease.

In recent years, the number of multistate outbreaks has increased, often through the contamination of foodstuffs during large-scale production at a single plant, as in a 2007 US outbreak involving peanut products and encompassing over 700 cases in 48 states. Efficient interstate and international distribution systems that deliver large amounts of the contaminated food over a wide area facilitate spread. Under these conditions, an attack rate as low as 0.5% can still produce many infections because of the large number of persons at risk. It is of concern that relatively small numbers of cases sprinkled over a massive area will be missed by local surveillance systems crippled by budgetary cutbacks.

PATHOGENESIS

Ingested S enterica cells that pass the stomach acid and swim through the intestinal mucous layer eventually reach the small bowel. Though it is not clear whether the initial contact there is with M cells, enterocytes, or both, the initial adherence is probably mediated by pili. Upon engagement of S enterica injection (type III) secretion systems, formation of membrane “ruffles” dramatically alters the normal host cell architecture within minutes (Figure 33–9). These ruffles are specialized plasma membrane sites of filamentous actin cytoskeletal rearrangement, normally induced by physiologic molecules such as growth factors. The bacterial secretion systems inject multiple other effectors coded by genes located within PAIs in the Salmonella genome. The virulence factors coded by PAI genes are either components of the injection system apparatus itself or the effector proteins it injects.

The ruffles seem to engulf the organism in an endocytotic vacuole and allow it to transcytose from the apical surface to the basolateral membrane. Once in the cell, S enterica multiplies in a vacuole and continues on through the cell, entering the lamina propria. There the bacteria typically induce a profound inflammatory response, and are phagocytosed by neutrophils and macrophages. Then, by deploying a second injection secretion system from inside the host macrophage, bacteria induce apoptosis, killing the macrophages and thus persisting in the lamina propria. This process contrasts with Shigella, which escapes the endocytotic vacuole (and double vacuole) to the cytoplasm and prefers to invade adjacent enterocytes rather than move through to the submucosa. Alternatively, it has recently been demonstrated that Salmonella may become nonreplicating persisters within host cells, according to survival programs induced by vacuolar acidification and nutritional deprivation. Subsequent resumption of intracellular growth by these persisters may be the basis for relapsing infection.

Although some enterotoxins have been described in Salmonella, their role in diarrhea is unclear. The best estimate is that the invasion and transcytosis of enterocytes together with the associated increased vascular permeability and inflammatory response are enough to account for the diarrhea. The release of prostaglandins and chemotactic factors may trigger inflammation and biochemical changes in enterocytes. Although the process remains localized to the mucosa and submucosa with most S enterica strains, some invade more deeply, reaching the bloodstream and distant organs. Some serotypes (eg, S ser. Choleraesuis) even invade so rapidly that they produce minimal diarrhea and are isolated more frequently from the blood than stool.

IMMUNITY

Evidence that both humoral and cell-mediated immune responses are stimulated by infection with S enterica is ample. Salmonella pathogens are also able to exploit some of these responses to compete successfully with resident microbiota and promote their own survival within the host. While several of these host–pathogen interactions have been characterized in detail, the key determinants in the outcome of infectious episodes (resolution, dissemination, chronic infection) remain to be determined.

ENTERIC (TYPHOID) FEVER (SALMONELLA SEROVAR TYPHI)

EPIDEMIOLOGY

Typhoid fever is a strictly human disease. Chronic carriers of serovar Typhi are the primary reservoir. Some patients become chronic carriers for years (hence the infamous “Typhoid Mary” Mallon), usually because of chronic infection of the biliary tract when gallstones are present. All cases can and should be traced back to their human source. If a patient with typhoid has not traveled to an endemic area, the source must be a visitor or someone else who prepared food. The pathogen can be transmitted in the water supply in developing endemic areas or where defects in any system allow sewage from carriers to contaminate drinking water. Transmission is by the fecal–oral route. The infecting dose of 10⁵ to 10⁶ bacteria is intermediate between Shigella and most S enterica, and decreases in the presence of the capsular Vi antigen. Three serotypes called Paratyphi A, B, and C have features similar to S Typhi, including the production of an enteric fever syndrome; cases are likewise traceable to a human source.

Typhoid fever is still an important cause of morbidity and mortality worldwide with 16 million cases and 600 000 deaths a year. In developed countries, it is mostly seen in travelers returning from endemic areas such as Latin America, Asia, and India. Visitors from these areas who are carriers are often the source of isolated cases. The decline in disease in industrialized nations largely reflects the availability of clean water supplies and improved disposal of fecal waste.

PATHOGENESIS

There is no animal model for the strictly human S Typhi. The details of the cellular events are inferred from studies of Typhimurium, which in mice produces a disease similar to typhoid (thus the name). The invasion and killing of intestinal M cells and macrophages are presumed to follow the same pattern as that of S enterica. Two differences are the Vi surface polysaccharide and the extended multiplication of Typhi in macrophages. In the submucosa, Vi (for virulence) retards neutrophil phagocytosis by interfering with complement deposition in a manner similar to that of other bacterial surface polysaccharides. This may favor uptake by macrophages, where at least some Typhi cells establish a privileged niche and the Vi⁺ phenotype favors intracellular multiplication. Like other serotypes of Salmonella, Typhi remains within a membrane-bound vacuole, but unlike them, rather than killing the macrophage, it enters a stage of extended replication.

The primary difference between Typhi and the other serotypes is the prolonged intracellular survival in macrophages. This is due to Typhi’s ability to inhibit the oxidative metabolic burst and continue to multiply. As the bacteria proliferate in macrophages, they are carried through the lymphatic circulation to the mesenteric nodes, spleen, liver, and bone marrow, all elements of the reticuloendothelial system (RES). At the RES sites, Typhi continues to multiply, infecting new host macrophages. Rather than the acute inflammatory response seen with S enterica, S Typhi generates a mononuclear response and often not enough irritation to cause diarrhea. This may be due to the downregulation of innate toll-like receptor responses in the intestinal mucosa by the Vi antigen.

Eventually, the increasing bacterial population begins to reach the bloodstream (Figure 33–8). The entry of gram-negative bacteria and their LPS endotoxin into the blood starts the fever, which slowly increases and persists with the continued seeding of S Typhi. This sometimes results in metastatic infection of other organs including the urinary tract and the biliary tree; the latter causes reinfection of the bowel. This cycle, beginning and ending in the small intestine, takes approximately 2 weeks to complete.

IMMUNITY

Natural infection with S Typhi confers immunity, and reinfection is rare unless the course was shortened by early administration of antimicrobials. The immune response is both T_H1 and T_H2-mediated. In nonfatal cases, antibody and activated macrophages eventually subdue the untreated infection over a period of about 3 weeks. Which antigens stimulate this immunity is not clearly understood. The Vi antigen is usually credited, but various surface proteins are also candidates.

SALMONELLOSIS: CLINICAL ASPECTS

MANIFESTATIONS

The clinical patterns of salmonellosis can be divided into gastroenteritis, bacteremia with and without focal extraintestinal infection, enteric fever, and the asymptomatic carrier state. Any Salmonella serotype can probably cause any of these clinical manifestations under appropriate conditions, but in practice the S enterica serotypes are associated primarily with gastroenteritis, while Typhi and related serotypes (Paratyphi) cause enteric fever.

Gastroenteritis

Typically, the episode begins 24 to 48 hours after ingestion, with nausea and vomiting followed by, or concomitant with, abdominal cramps and diarrhea. Diarrhea persists as the predominant symptom for 3 to 4 days and usually resolves spontaneously within 7 days. Fever (39°C) is present in about 50% of the patients. The spectrum of disease ranges from a few loose stools to a severe dysentery-like syndrome.

Bacteremia and Metastatic Infection

The acute gastroenteritis caused by S enterica can be associated with transient or persistent bacteremia. Frank sepsis is uncommon, except in those with compromised cell-mediated immunity. Salmonella infection in patients with acquired immunodeficiency syndrome (AIDS) is common and often severe. Bacteremia occurs in 70% of these patients and can cause septic shock and death. Despite adequate antimicrobial coverage, relapses are common. Patients with lymphoproliferative disease, perhaps owing to T-cell defects similar to those in patients with AIDS, are also highly susceptible to disseminated salmonellosis. Metastatic spread by salmonellae is a significant risk when bacteremia occurs. These organisms have a unique ability to colonize sites of preexisting structural abnormality, including atherosclerotic plaques, sites of malignancy, and the meninges (especially in infants). Salmonella infection of the bone typically involves the long bones; in particular, sites of trauma, sickle cell injury, and skeletal prostheses are at risk.

Enteric Fever

Enteric fever is a multiorgan Salmonella infection characterized by prolonged fever, sustained bacteremia, and profound involvement of the mesenteric lymph nodes, liver, and spleen. The manifestations of typhoid (Figure 33–10) have been well documented in human volunteer studies conducted during vaccine trials. The mean incubation period is 13 days, and the first sign of disease is fever associated with a headache.

The fever rises in a stepwise fashion over the next 72 hours. A relatively slow pulse is characteristic and out of character with the elevated temperature. In untreated patients, the elevated temperature persists for weeks. A faint rash (rose spots) appears during the first few days on the abdomen and chest. Few in number, these spots are readily overlooked, especially in dark-skinned persons. Many patients are constipated, although perhaps one-third of patients have mild diarrhea. As the untreated disease progresses, an increasing number of patients complain of diarrhea.

Obviously, prolonged infection of the bloodstream is serious, and the effects of endotoxin can lead to myocarditis, encephalopathy, or intravascular coagulation. Moreover, the persistent bacteremia can lead to infection at other sites. Of particular importance is the biliary tree, with reinfection of the intestinal tract and diarrhea late in the disease. Urinary tract infection and metastatic lesions in bone, joint, liver, and meninges may also occur. However, the most important complication of typhoid fever is intra-abdominal hemorrhage, resulting from perforations through the wall of the terminal ileum or proximal colon at the site of necrotic Peyer patches. These occur in patients whose disease has been progressing for 2 weeks or more.

DIAGNOSIS

Culture of Salmonella from the blood or feces is the primary diagnostic method. Early in the course of enteric fever, blood is far more likely to give a positive culture result than culture from any other site. The media used for stool culture are the same as those used for Shigella. Failure to ferment lactose and the production of hydrogen sulfides (H₂S) from sulfur-containing amino acids are characteristic features used to identify suspect colonies on the selective isolation media. Characteristics of biochemical tests are used to identify the genus, and O serogroup antisera are available in larger laboratories for confirmation. Typhi has a pattern of biochemical reactions that is sufficient to characterize it without reference to its serogroup (D). All isolates should be referred to public health laboratories for confirmation and epidemiologic tracing. Serologic tests are no longer used for diagnosis.

TREATMENT

The primary therapeutic approach to Salmonella gastroenteritis consists of fluid and electrolyte replacement and the control of nausea and vomiting. Antibiotic therapy is usually not appropriate because it has a tendency to increase the duration and frequency of the carrier state. When used to eradicate the carrier state, it meets with erratic success and usually fails in the presence of coexisting biliary tract disease. Therefore, the use of antimicrobial agents in S enterica gastroenteritis is restricted to those with severe infections or underlying risk factors, particularly children. In these instances, antimicrobials are viewed as a measure to prevent systemic spread.

Antimicrobial therapy is clearly indicated in typhoid fever. Chloramphenicol and then ampicillin were the first antibiotics used and reduced the mortality rate from 20% to less than 2%. Use of ampicillin is now limited by widespread resistance, leaving the extended-spectrum cephalosporins (ceftriaxone, cefixime) and ciprofloxacin as preferred first-line agents. As seen in other pathogenic species (E coli, K pneumoniae), global patterns of antimicrobial resistance in S Typhi have been reshaped by the recent emergence of a single multidrug-resistant clone, known as H58. Nonetheless, with effective antimicrobial therapy, patients feel better in 24 to 48 hours, their temperature returns to normal in 3 to 5 days, and they are generally well in 10 to 14 days.

PREVENTION

Killed whole bacterial vaccines have been available for typhoid since the late 19th century, with protection in the range of 50% to 70%. Newer vaccines include one that uses a live attenuated Typhi strain and a polysaccharide vaccine containing the Vi antigen. The newer vaccines give slightly higher protection, but none gives protection lasting more than a few years. The newest vaccine contains Vi antigen conjugated to a bacterial protein in the manner of Hib, meningococcal, and pneumococcal vaccines. It shows promise for both higher efficacy and use in children less than 5 years of age. No human vaccine is available for the other Salmonella serotypes. When all is said and done, the provision of clean water supplies and the treatment of carriers would still go a long way toward eradicating typhoid. The importance of carriers and sanitation was emphasized by a 1973 typhoid outbreak among migrant workers in Florida. The source was traced to leakage of sewage into the water supply, failure of chlorination, and a chronic carrier. All three are required to sustain an outbreak when adequate sanitary infrastructure is in place.

KEY CONCLUSIONS

BACTERIOLOGY

Morphologically, Yersinia tend to be coccobacillary and to retain staining at the ends of the cells (bipolar staining). Growth and metabolic characteristics are the same as those of other Enterobacteriaceae, although some strains grow more slowly or have optimal growth temperatures lower than 37°C. The genus includes 11 species, of which Yersinia pestis, Yersinia pseudotuberculosis, and Yersinia enterocolitica are the pathogens for humans. Yersinia pestis is antigenically homogenous, but Y pseudotuberculosis and Y enterocolitica have multiple O and H antigen serotypes. Yersinia are primarily animal pathogens, with occasional transmission to humans through direct or indirect contact. Yersinia pestis, the cause of plague, is discussed in Chapter 36, although features of its pathogenesis common to other Yersinia are included in the following discussion.

YERSINIA DISEASES (Y PSEUDOTUBERCULOSIS AND Y ENTEROCOLITICA)

EPIDEMIOLOGY

In animals, Y pseudotuberculosis causes pseudotuberculosis, a disease characterized by local necrosis and granulomatous inflammation in the lymph nodes, spleen, and liver. The portal of entry for humans is the gastrointestinal tract, presumably by consumption of contaminated food or water. In most cases, animals, including wild animals, are the most likely source of infection, but the exact mode of transmission is unknown. Geographic variation in the frequency of Y enterocolitica infections is marked. The highest rates are reported from Scandinavian and other European countries, with much lower rates in the United Kingdom and the United States. Low isolation rates may be partially attributable to the difficulty of isolating Y enterocolitica from stool specimens.

PATHOGENESIS

Enteropathogenic Yersinia enter the human host in contaminated food and invade the M cells of the Peyer patch. The invasive process and its effect on the host cell are driven by a large array of virulence factors that are deployed under complex genetic and environmental regulation. These proteins include invasin, which binds to integrins on the surface of host cells, and the major effector proteins called Yersinia outer membrane proteins (Yops). The Yops are delivered by yet another injection (type III) secretion system. When injected into the host cell, they trigger cytotoxic events, including disruption of biochemical pathways (dephosphorylation, serine kinase), sensor functions, and the actin cytoskeleton.

Some of the virulence factors produced by Yersinia are regulated in response to either temperature or free calcium (Ca²⁺) concentration. The physiologic temperature in a mammalian host is different from that in an insect or the environment, and the intracellular calcium concentration is markedly different from that of extracellular fluids. By sensing the environment, Yersinia is able to express or suppress virulence factors at different stages of the pathogenic process. The results seem timed to support the pathogenic strategy of Yersinia, which is to paralyze the phagocytic activity of defending macrophages and neutrophils and thus nullify the host cellular immune response. The virulence determinants are encoded both on the bacterial chromosome and on a plasmid that contains genes for the secretion apparatus and the Yops. Another genetic component is a PAI, which is found only in the three pathogenic species and not other Yersinia.

The biological outcome of this extraordinary multifactorial process is the enhanced capacity of the pathogenic Yersinia to enter and replicate within the RES and to delay the cellular immune response. This leads to the formation of microabscesses and destruction of the cytoarchitecture of Peyer patches and the mesenteric lymph nodes. The systemic symptoms seen with dissemination can largely be attributed to the effects of endotoxin.

Yersinia pestis is a specialized variant closely related to Y pseudotuberculosis. Instead of entering the intestinal tract, Y pestis reaches the dermal lymphatics by the bite of an infected flea. It has its own adhesin similar to invasin and two plasmids not found in the enteropathogenic Yersinia. Unique virulence factors for Y pestis include a capsular protein antigen with antiphagocytic properties, a plasminogen activator protease that promotes adherence to basement membranes, and a fibrinolysin that may play a survival role in the flea.

YERSINIA INFECTIONS: CLINICAL ASPECTS

Both Y enterocolitica and Y pseudotuberculosis cause acute mesenteric lymphadenitis, a syndrome involving fever and abdominal pain that often mimics acute appendicitis. Yersinia enterocolitica produces a wider variety of manifestations as well. The most common of these is enterocolitis, which usually occurs in children. It is characterized by fever, diarrhea, and abdominal pain. It also causes enteric fever, terminal ileitis, and a polyarthritic syndrome associated with its diarrheal manifestations. Few laboratories in the United States routinely screen stools for Yersinia because yield has been low and good selective media are not available.

The role of antimicrobial therapy in enteric Yersinia infections is uncertain, because they are usually self-limiting. Yersinia pseudotuberculosis is susceptible to ampicillin, cephalosporins, aminoglycosides, and tetracyclines, but Y enterocolitica is usually resistant to penicillins and cephalosporins through the production of β-lactamases.

All the Enterobacteriaceae are capable of producing opportunistic infections of the type discussed under E coli. None is considered a proven cause of enteric disease, although no doubt some will be in the future. The genera isolated in at least moderate frequency are discussed briefly as follows. There are many other less common species.

KLEBSIELLA

The most distinctive bacteriologic features of the genus Klebsiella are the absence of motility and the presence of a polysaccharide capsule. This gives colonies a glistening, mucoid character and forms the basis of a serotyping system. Over 70 capsular types have been defined, including some that cross-react with those of other encapsulated pathogens, such as Streptococcus pneumoniae and Haemophilus influenzae. Limited studies suggest that the capsule interferes with complement activation in a way similar to the other encapsulated pathogens. Klebsiella also express several types of pili on the cell surface which probably aid in adherence to respiratory and urinary epithelium.

Klebsiella pneumoniae, the most common species, is able to cause classic lobar pneumonia, a characteristic of other encapsulated bacteria; most Klebsiella pneumonias are indistinguishable from those produced by other members of the Enterobacteriaceae. Highly mucoid K pneumoniae bearing K1 or K2 capsule have been associated with distinctive clinical syndromes featuring liver abscess and endophthalmitis, particularly in southeast Asian countries. Of all the Enterobacteriaceae, Klebsiella species are now among the most resistant to antimicrobial agents. In the early 2000s, a single K pneumoniae clone (Sequence Type 258), now notorious for its near pan-resistant properties, emerged as a devastating cause of hospital-acquired infection (bloodstream, respiratory tract, urinary tract) in the northeastern United States and spread rapidly around the globe. The rise and spread of this and other highly-resistant “superbugs” has led to concerns about a post-antibiotic era.

ENTEROBACTER

Enterobacter species generally ferment lactose promptly and produce colonies similar to those of Klebsiella, though not as mucoid. A differential feature is motility by peritrichous flagella, which are generally present in Enterobacter species but uniformly absent in Klebsiella. Enterobacter species, which are generally less virulent than Klebsiella, have attracted increasing attention as a cause of infections acquired in the hospital, where their intrinsic antibiotic resistance properties undoubtedly confer a selective advantage. In addition to ampicillin, most isolates are resistant to first-generation cephalosporins, but may be susceptible to later-generation cephalosporins. However, resistance to cephalosporins may emerge (via derepression of β-lactamase production) during antibiotic treatment of patients with infectious foci such as incompletely drained abscesses or devitalized/necrotic tissue, where bacteria may persist.

SERRATIA

Serratia strains ferment lactose slowly (3-4 days), if at all. Some produce distinctive brick-red colonies. Although less common, this genus produces the same range of opportunistic infections seen with other Enterobacteriaceae. Serratia strains show consistent intrinsic resistance to ampicillin and cephalothin/cefazolin, and with additional acquisition of resistance plasmids, to many other antimicrobials including the aminoglycosides. Sporadic infections and nosocomial outbreaks with multiresistant strains have often been difficult to control.

CITROBACTER

The genus Citrobacter, though biochemically and serologically similar to Salmonella, is an uncommon cause of opportunistic infection. Like many other Enterobacteriaceae, Citrobacter strains may be present in the normal intestinal flora and cause opportunistic infections. Despite reports of association with diarrheal disease, present evidence does not indicate that Citrobacter should be considered an enteric pathogen of humans. Citrobacter freundii has been associated with neonatal meningitis and brain abscess.

PROTEUS, PROVIDENCIA, AND MORGANELLA

Proteus, Morganella, and Providencia are also opportunistic pathogens found with varying frequencies in the normal intestinal flora. Proteus mirabilis, the most commonly isolated member of the group, is one of the most susceptible of the Enterobacteriaceae to the penicillins; this characteristic includes moderate susceptibility to penicillin G. Other Proteae (typically, indole-positive species like Proteus vulgaris) are intrinsically resistant to ampicillin and the cephalosporins. Proteus mirabilis and P vulgaris share the ability to swarm over the surface of media, rather than remaining confined to discrete colonies. This characteristic makes them readily recognizable in the laboratory—often with dismay because the spreading growth covers other organisms in the culture and thus delays their isolation. Swarming along with motility could facilitate the production of UTI by propelling Proteus up urinary catheters. Proteus and Morganella differ from other Enterobacteriaceae in the production of a very potent urease, which aids their rapid identification. It also contributes to the formation of urinary stones and produces alkalinity and an ammoniac odor to the urine. Providencia species do not produce urease, are the least frequently isolated, and are generally the most resistant of the group to antimicrobials.