Nongrowing bacteria evade the bactericidal activity of beta-lactam antibiotics. We sought to determine if slow growth rate also alters bactericidal activity. The bactericidal activity of two beta-lactams on Escherichia coli grown in glucose limited chemostats was compared for generation times ranging from 0.7 to 12 h. The degree of killing varied with drug structure and with E. coli strain. However, all killing rates were a constant function of the bacterial generation time: slowly growing bacteria became progressively more phenotypically tolerant to beta-lactam antibiotics as the generation time was extended.

Pneumococcal cell wall induces meningeal inflammation in rabbits injected intracisternally with >10 5 cell equivalents. Both of the major cell wall components, teichoic acid and peptidoglycan, contribute to this inflammatory activity although responses differ depending on the chemical nature, size, and complexity of these fractions. Challenge with teichoic acid (membrane or wall associated) results in greater inflammation at 5 hr than at 24 hr. Degraded teichoic acid is inactive. In contrast, the inflammation caused by whole cell wall or high-molecular-weight peptidoglycan-containing fractions increases in intensity from 5 to 24 hr. Peptidoglycan fractions lose activity at 24 hr when hydrolyzed to disaccharide-stem peptide moieties. Generation of free cellwallcomponents in cerebrospinal fluid as, for example, during treatment with antibiotics that are bacteriolytic as well as bactericidal, could contribute to increased inflammation in the subarachnoid space.The cell wall of pneumococci, located under a layer of capsular polysaccharide, remains surprisingly accessible to and reactive with the host environment [1]. We have shown that meningeal inflammation in rabbits is induced when whole pneumococci, with or without capsular polysaccharide, reach a density of >10 5 cfu/ml of CSF. This inflammatory response is remarkably similar to the inflammation following challenge with 105 cellequivalents of isolated cell wall, but not of isolated capsule [2]. Because the cell wall is a complex macromolecule with many possible sites of interaction with several host defense systems, the identification of which cell wall component(s) is active in inducing inflammation during pneumococcal meningitis is of considerable importance.The pneumococcal cell wall is composed of two major polymers: a peptidoglycan and a ribitolphosphate teichoic acid of unusually complex structure that contains phosphorylcholine [3]. The inter-

The relative contribution of bacterial components to the induction of inflammation during Streptococcus pneumoniae meningitis is unknown. Several strains of pneumococci with differences in cell surface characteristics (capsule or cell wall) were compared for the effect on the inflammatory response evoked during infection of the cerebrospinal fluid (CSF) in vivo. The presence of bacterial capsular polysaccharide was not necessary for bacterial growth in CSF in vivo but correlated with greater CSF bacterial density. CSF inflammatory changes began to appear when the bacterial concentration exceeded 10 5 cfu/ml, regardless of the pneumococcal strain. CSF inflammatory changes could be invoked by cisternal instillation of 10 5-106 cell equivalents of whole, heat-killed unencapsulated strains or their isolated cell walls but not by similar concentrations of heat-killed encapsulated strains or isolated capsular polysaccharide. Hypoglycorrhachia was observed only during inflammation caused by live bacteria. The inflammatory response characteristic of naturally acquired pneumococcal meningitis can be reproduced by challenge with both encapsulated and unencapsulated bacteria. The bacterial cell wall appears to be the most potent pneumococcal surface component in inducing CSF inflammation.

The bactericidal activity of 23 I-lactam antibiotics was compared in slowly growing bacteria cultured in a chemostat. In an attempt to mimic possible in vivo conditions, slowly growing cultures were produced by limitation of iron, glucose, phosphate, or magnesium. Only select antibiotics remained effectively bactericidal against slowly growing cells. For these compounds, the rate of antibiotic-induced loss of viability was a constant when killing was expressed per generation (in contrast to absolute time) in that slowly growing bacteria were killed proportionately more slowly. Individual antibiotics differed greatly, however, in their specific bactericidal activities against slowly growing cells, i.e., in the absolute degree of killing elicited during exposure of the bacteria to MIC equivalents of the drugs. Specific bactericidal activities varied not only with drug structure but also with the bacterial strains and, to a lesser extent, with the nature of the growth-limiting nutrient. In slowly growing cultures exposure to the low drug concentrations studied here (near MIC) caused killing without detectable lysis. Antibiotics with high specific bactericidal activities were capable of rapidly killing cultures of slowly growing pathogens despite extremely long generation times approaching those reported for in vivo growth rates.The in vitro bactericidal activity of an antibiotic does not always correlate with therapeutic efficacy. The reasons for this lack of correlation are likely to be multiple (35). One explanation which has received only recent attention is the probable phenotypic differences between bacteria growing in the nutritionally limited in vivo environment and those growing under optimum conditions in the laboratory. Phenotypically induced changes in microorganisms have been shown to alter markedly their susceptibility to the bactericidal activity of antimicrobial agents; these changes include alterations in drug permeability across the outer or cytoplasmic membranes (3,7,14), alterations in the structure of peptidoglycan such that it resists degradative enzymes (17,23,26,28), or alteration of penicillin-binding proteins (4,9,13,25).A slow growth rate (12,22,24) and a restricted availability of iron (5, 33) and possibly other nutrients appear to be characteristic of many infections in vivo. These two parameters could be major contributors to phenotypic tolerance (i.e., relative insensitivity to the bactericidal effect of drugs) in vivo (E. Tuomanen, Rev. Infect. Dis., in press). Since the earliest days of the antibiotic era it has been recognized that slowly growing bacteria are less susceptible to antibiotic action than those growing at optimum rates (16).The chemostat has already proved useful in studies of the effect of growth rate and nutritional status of microorganisms on their susceptibility to antimicrobial agents (8,14,19,20,29). The purpose of the study described here was to compare the bactericidal activity of a range of 1-lactam antibiotics on several species of bacteria growing at reduced ...

No abstract