What causes antibiotic resistance?
Antibiotic Resistance – Bells Are Tolling!
Antibiotic resistance
Ever since the introduction of antibiotics, bacterial resistance has been an issue; however, it is not until recently that this issue has become one of major importance, attracting the attention of the World Health Organization, policymakers, and politicians. The essential reason for the change is that we are exhausting our arsenal of new antibiotics. Many pharmaceutical companies are now reluctant to invest in new antibiotics and the difficulties in discovering new antibacterial drugs that are safe enough to be given to clinical patients are immense. While there is no shortage of antibacterial compounds, the vast majority of them would fail current safety standards. The problem was perceived to be driven by the Gram-positive bacteria, such as MRSA, but actually the acute crisis is the continued capability to treat infections caused by Gram-negative bacteria.
Mutations in the bacterial chromosome
Bacterial infections are composed of, literally, millions of individual cells. These may be in a liquid environment such as the blood or urine or, perhaps, more likely they are congregated at a focus of infection. These bacteria are not really individual cells but, as we have seen, they form a colony of cells with, largely, an identical set of genes. The survival of the colony requires cooperation between the individual cells and the ‘sacrifice’ of the majority so that the ‘gene pool’ can continue to exist. The simplest form of resistance comes from the ability of mutations to occur within genes. This has an average frequency of approximately 1 in 10 (7) bacterial divisions for any single gene. Thus within one billion bacteria, which make such a colony, there might be a hundred bacteria that have undergone spontaneous mutation in a specific gene and this mutation may confer resistance to a specific antibiotic. When this population is treated with the antibiotic, the vast majority of the bacteria, all but the hundred with the mutation, are inhibited or killed but the mutants are able to prosper and they start dividing. They can soon replace the original infection site with their progeny. Mutations occur during DNA replication prior to bacterial division, which means it is a symbiotic relationship theoretically possible to have a population of bacteria with no resistant mutants when antibiotic treatment begins; however, if the concentration of antibiotic is insufficient, DNA replication will continue and mutants can form. This type of mutational resistance is the reason why doses of antibiotic have to exceed a concentration threshold to prevent the mutants surviving; this has more recently been described as the ‘Mutant Prevention Concentration’.
Once a single mutation has occurred and the resistant bacteria have taken over the infection site, it is possible for a second mutation to occur by the same process, giving an even higher degree of resistance, and the progeny of this double mutant will divide and then take over the population. The gene encoding the target for a given antibiotic is usually the gene in which the mutations take place; for instance, the gyrA gene encodes a subunit of the enzyme DNA gyrase to which the antibacterial drug ciprofloxacin binds. Sequential mutations in the gyrA gene produce a DNA gyrase molecule that binds ciprofloxacin far less tightly and thus confers resistance. The original gyrA gene encodes a protein with a structure that has evolved over an aeon and presumably this has resulted in the most efficient structure for the role it performs. While the mutations still allow the original role to occur, it may well do so less efficiently than the non-mutated protein. This means that there is an efficiency ‘cost’ for the bacterium to carry the mutation and often these bacteria divide more slowly in the absence of antibiotics than the bacteria without the mutant. This result in pressure, once the antibiotic has been removed, for reverse or back mutations to occur and this is often seen. This is a reason why strategies to overcome bacterial resistance have suggested that the restriction or banning of certain antibiotics could reverse the development of resistance; however, back mutations are not the only mutations that can occur under these circumstances. Mutations can occur in other sites in the genome, which improve the division rate; these are called compensatory mutations. The bacterium retains its resistance mutations and is also able to decrease the ‘cost’ that these mutations have caused.
Some strains of bacteria are able to pass into a transient hypermutator state. Effectively this means that mutations will occur at a much higher frequency than 1 in 107 bacterial divisions, maybe an increase of 1,000-fold. Whereas this could be an advantage in a situation where the bacteria are likely to be challenged with antibiotics, this mutation rate will occur in all genes producing disadvantageous mutations at least as often as those giving an advantage. Disadvantageous mutations are almost always lethal, so the hypermutator state is often precarious for a bacterial population and they often pass back into their normal state. The capability of increased frequency of mutation is found as a permanent feature in some strains of bacteria that have become resistant to many antibiotics and is thought to have been a contributory factor to their multi-resistant character. Although mutation has been crucial to the development of all resistance mechanisms at some point, mutations emerging in the chromosomes of bacteria treated with antibiotics are a surprisingly rare form of resistance.
Plasmids and mobile structures
The Japanese were often clustered together iH the first to notice that some bacteria became resistant not to just one antibiotic but to four at the same time. This could not be explained by chromosomal mutations as it would have required a mutation frequency of 1 in 1028 bacterial cell divisions, producing a mass of bacteria greater than that of the moon. Clearly this did not happen and the resistance genes were subsequently found to have been imported en bloc by a plasmid. These circular molecules of DNA are independent of the bacterial chromosome and have their own replication origin.
The plasmids conferring resistance in clinical bacteria usually replicate at the same time as the bacterial chromosome, so that the bacterium has approximately two to five copies of the plasmid. They have the genetic machinery that allows them to transfer from one bacterium to another; this is actually an extremely difficult process and requires approximately twenty-five genes but it allows the plasmid to pass from one strain to another and from one species to another, even travelling to quite different bacterial species. The constraint on unlimited transfers between bacteria is the defense mechanism that bacteria have evolved to protect themselves from incoming DNA. They produce endonucleases that cut and degrade incoming DNA and the plasmid has to overcome these restriction enzymes in order to survive in the new bacterium. The closer the new host is to the old one, the less restriction will occur. These endonucleases or restriction enzymes are able to recognize specific short sequences of DNA and, as a by-product of this capability; they are now a crucial element in many molecular biological procedures.
Chromosomal resistance genes are usually under a repression system so that they are not activated, whereas those on plasmids are usually maximally expressed to give the highest levels of resistance continuously. Re-examination of some of the earliest bacteria ever identified shows that the plasmids themselves have been present for a long time, though the presence of resistance genes is, for the most part, a recent development. The genes conferring resistance are often clustered together on plasmids, indicating that there is a preferred location. This gives a clue as to how they were captured by the plasmid. Bacterial DNA is littered with elements known as insertion sequences. These are relatively short, about 700–2500 nucleotide base pairs long, mobile elements of DNA that can move from one area of DNA to another. At the extremities of the insertion sequence are often inverted repeats and between these there are often genes that encode enzymes facilitating movement of the element.
Similar insertion sequences can often align themselves closely together on the same sequence of DNA. When this happens, the whole region of DNA, including the two insertion sequences and the region in between, can be moved. If there are other genes, in this case resistance genes, between the insertion sequences, this composite element is known as a transposon.
Often transposons insert into specific sites within DNA and this is commonly the case in plasmids. By the successive introduction of resistance genes, sited in transposons, into a particular location within the plasmid DNA, a cluster of resistance genes form. Often the structure of the original insertion sequences of the early transposons is disrupted by further transposition events, bringing in new resistance genes. This fixes the early resistance genes in the plasmids and prevents them being transposed ‘out of’ the plasmids. Insertion sequences can, in theory, surround and pick up any gene in a transposon; however, selection plays an important role and the reason why we find clusters of resistance genes is because antibiotic usage has selected the plasmids that have particular resistance gene combinations. Transposons along with plasmids provide a symbiotic relationship often clustered together iH with the host bacteria. They often carry accessory genes, resistance in this case, that are beneficial but not usually essential for the survival of the organisms. If they are lost, the bacterial population could still survive. Often, however, the exposure to antibiotics is so great that it is more beneficial for the bacterium to carry the resistance gene in the chromosome than on a plasmid.
Transposons carrying resistance genes are increasingly found located in the bacterial chromosome. It could be argued that this is not the evolutionary end point for resistance genes because many multiresistant bacteria that are spreading clonally through the clinical populations retain their resistance genes on identical plasmids. It is likely that there is a ‘match’ between certain plasmids and bacteria.
Plasmids, as extra-chromosomal DNA, do provide an energy drain on the resources of the cell. In some successful cases, where a plasmid has established itself in a particular pathogen, the cost has been ameliorated by mutations in the plasmid and the host DNA, ensuring that this becomes a favoured combination. There is a further advantage for the plasmid as not every bacterial cell within the population needs to carry the plasmid, though they often do, because if the bacteria are suddenly challenged with antibiotics then only those with the plasmid survive.
The capture of resistance genes by insertion sequences is a relatively inefficient process and does not alone explain the rapid acquisition of resistance genes by transposons and plasmids. Within each are often found integrons. These are short DNA sequences that can capture genes by encoding an integrase enzyme. The gene, int1, encoding integrase, is closely linked to an attachment site (att), into which captured genes are inserted, and a promoter, from which the inserted gene is expressed. The gene cassettes that integrons capture often do not confer resistance, but antibiotic usage ensures that many of those integrons identified in clinical bacteria do carry resistance genes.
Mechanisms of resistance
Mutation of the target of an antibiotic is not an option for plasmid-mediated resistance. The resistance mechanism has to be dominant and is usually, though not always, the result of a single gene. Almost all bacteria possess genes encoding enzymes that export often unwanted metabolites. When encoded by the chromosome, these efflux pumps are usually under strict repressed control and are activated only when needed. Some of them are able to efflux antibiotics, thus providing a low level of resistance as they export the drug faster than it can enter. Some plasmids have acquired genes for efflux and are able to confer resistance to antibiotics such as tetracycline. The efficiency is dependent on the concentration of antibiotic outside the cell and, as more is being imported, the efflux pump is soon overwhelmed. Thus they give relatively low levels of resistance and are, in many cases, just the first line of defence. Clinical concentrations of antibiotics are high and the resistance mechanisms must match this. The most widespread mechanism is inactivation of the antibiotic and this accounts for the vast majority of resistance genes identified. This is manifested in two ways: by the enzymic destruction of the antibiotic or by the addition of a chemical group that is added to the antibiotic so that it is no longer taken up or recognized as an active antibiotic.
Enzymic destruction is exemplified by just one enzyme, the beta-lactamase. Many bacteria have genes encoding them on their bacteria chromosome but in clinical bacteria, the infiltration of betalactamases has been on plasmids. There are now more than 1,000 identified in clinical bacteria. The enzyme attacks the beta-lactam ring, which is crucial to the function of penicillin’s, cephalosporins, and carbapenems. Beta-lactamases are a prime example of convergent evolution, as four distinct molecular classes, A–D, have been identified. Class A is the chromosomal beta-lactamases of Grampositive bacteria and has evolved primarily to confer resistance to penicillin’s. Class C is the chromosomal beta-lactamases from Gram-negative bacteria and its primary target are the cephalosporins. The class D beta-lactamases are chromosomal beta-lactamases of one genus of Gramnegative bacteria, Acinetobacter. All three classes have the same basic mechanism of action, using a serine residue for their enzyme catalysis, though their overall structures are completely different.
Class B beta-lactamases have a different mechanism of action; they are metallo-enzymes, so use a zinc ion to catalyse the destruction of the antibiotic and they are particularly effective against the carbapenems. The origins of this class of beta-lactamase are unknown. The class A beta-lactamases were transported from their Gram-positive origins to the plasmids of Gram-negative bacteria. The most prevalent is called TEM-1 after the little girl, Temoira, who provided the bacteria from which it was first identified. This beta-lactamase was unknown before the introduction of ampicillin and has now become the most widespread antibiotic resistance mechanism in the world, so much so that a quarter of the population carry commensal bacteria containing this resistance gene. The TEM-1 beta-lactamase provided a very efficient resistance to the penicillin’s but was largely unaffected by the cephalosporins; therefore there was a massive rise in the number of new cephalosporins capable of overcoming this resistance. However, in 1982 there was an outbreak in a neonatal unit in Liverpool, England. The causative bacterium, Klebsiella oxytoca, was treated with the cephalosporin ceftazidime until the bacterium became resistant. The TEM-1 beta-lactamase that it was carrying had mutated at position 164 with the effect of opening the active site and allowing ceftazidime to be destroyed. Once this first mutation had taken place other mutations soon followed, until well over a hundred different enzymes all derived from the original TEM molecules have been found and virtually all cephalosporins could now be resisted. These were known as extended-spectrum beta-lactamases, or ESBLs. A Class A beta-lactamase similar to TEM though far less common is SHV-1; it has the same activity against penicillin’s. The SHV-1 beta-lactamase also started to mutate until it also produced over a hundred ESBLs. This had the effect of not only reducing the usage of cephalosporins but also restricting their further development as new antibiotics.
Actually the proportion of ESBLs based on TEM and SHV are now declining, as a new group of Class A ESBLs is emerging. These are the CTX-M ESBLs, CTX standing for cefotaxime, the cephalosporin antibiotic that is the primary target of these enzymes. Unlike the plasmid origin of TEM and SHV, the CTX-M enzymes derive from the chromosomal beta-lactamases of Kluyvera species, which are closely related to Escherichia coli. The CTX-M-3 beta-lactamase is identical in structure to the chromosomal beta-lactamase in Kluyvera ascobata and appears to have been transferred directly from that species via a transposon and plasmid. CTX-M-3 is relatively rare in clinical bacteria but a single mutation at ensure that the HKluyvera appears to have happened on five occasions and has provided five distinct genetic subgroups of CTX-M beta-lactamases, each group named after the first enzyme discovered. The CTX-M-1 subgroup, which includes CTX-M-3 and CTX-M-15, was derived from Kluyvera ascorbata; the CTX-M-2 subgroup from a different strain of Kluyvera ascorbata. The CTXM-8 and CTX-M-9 subgroups were derived from two distinct strains of Kluyvera georgiana. The species that is progenitor of the last subgroup, CTX-M-25, is not known. The subgroups are not closely related to each other but the enzymes within each subgroup are very similar. Therefore, in each subgroup a single gene will have transferred into clinical bacteria from Kluyvera and then started to mutate; the selective pressure of different cephalosporin antibiotics in clinical use will have allowed a myriad of different CTX enzymes to emerge within each subgroup. After CTX-M-15, the most important of these is CTX-M-14, a member of the CTX-M-9 subgroup, as this has been the most prevalent ESBL in clinical bacteria from the Iberian Peninsula and, in the United Kingdom, in bacteria isolated from farm animals. This probably reflects the antibiotic usage in these areas.
The decline of the cephalosporins has produced a greater clinical reliance on the carbapenems. These antibiotics are, by and large, immune to the ESBLs, although there is a notable exception for strains containing the CTX-M-15 beta-lactamase, which can become carbapenem resistant as long as other resistance mechanisms, such as efflux pumps, are working in concert. Carbapenem resistance has come from new Class A beta-lactamases such as KPC-1 in Klebsiella pneumoniae. However, the largest group of plasmid-encoded beta-lactamases able to confer resistance to carbapenems comes from Class B. Two large groups of closely related beta-lactamases, IMP and VIM, have evolved over the last decade or so to confer carbapenem resistance. They are efficient but the greatest concern has been expressed about the emergence of another Class B beta-lactamase that was originally identified in New Delhi, India, called NDM-1. The gene encoding this beta-lactamase has the ability to migrate into many different bacterial species and can confer high-level carbapenem resistance.
The prototypes of the beta-lactamases found in clinical bacteria did not evolve entirely during the time that we started using antibiotics. They were initially defence mechanisms used by soil bacteria to gain an advantage against other bacteria and fungi such as Penicillium notatum. Without these enzymes, these bacteria would never have been able to survive. They have then been transported to clinical bacteria by successive, and hitherto unknown, interactions between environmental and clinical bacteria; in the latter, mutations have been selected by successive challenges with different antibiotics.
Many resistance mechanisms originally derive from the bacteria that originally produced the antibiotic. The antibiotic is used by the bacterium to destroy other bacteria in the immediate environment but it runs the risk of killing itself. Therefore, it had to evolve a defence system. An effective method is modification of the enzyme that produces the final antibiotic. If the gene is duplicated and then mutates, then the resultant enzyme may bind the antibiotic and attach a functional group to it to render it ineffective. This is the basis of aminoglycoside virus is being administered, aret modification. There are three enzymic groups: N-Acetyltransferases (AAC) catalysing acetyl CoA-dependent acetylation of an amino group; O-Adenyltransferases (ANT) catalysing ATP-dependent adenylation of a hydroxyl group; and O-Phosphotransferases (APH) catalysing ATP-dependent phosphorylation of a hydroxyl group effectively adding an acetyl, adenyl, or phosphate group respectively to various positions on the molecule, providing high-level resistance.
There are over fifty of these enzymes mainly encoded on plasmids. More recently, a new enzyme has emerged, AAC (6’)-ib-cr, which has the ability to add both an acetyl group and efflux fluoroquinolones – a very versatile enzyme!
Destruction and modification are the normal resistance mechanisms for antibiotics derived from natural sources; the bacteria face a greater challenge when the drug has no counterpart in nature. Trimethoprim and sulphonamides are completely synthetic and bacteria would never have previously been exposed to them, though resistance emerged within just a couple of years after the introduction of each compound. The initial resistance was caused by mutations in the target of the antibiotic, though this often only gave a low level of resistance. Bacteria that were highly resistant emerged soon afterwards and the genes responsible were found to have been introduced on plasmids. The mechanisms were different from those previously identified. The plasmids encoded an additional target enzyme: a dihydrofolate reductase in the case of trimethoprim resistance and a dihydropteroate synthetase for sulphonamide resistance. The additional target enzymes could still perform the task of the chromosomal enzyme but they were far less susceptible to inhibition by the antibiotics. Therefore they ‘bypass’ the antibiotic’s inhibition of the chromosomal enzyme. Although this can produce very high levels of resistance, an increase of up to 10,000-fold, it is a fairly inefficient method of resistance as the bacterium has to expend energy to produce two enzymes, one of which is effectively useless as it is being inhibited. This mechanism can only be employed when there are relatively few original enzyme molecules produced by the cell, up to just ten in these two cases. The origin of these enzymes is not clear. Some came from mutations in chromosomal enzymes of closely related species, though these tended to give low levels of resistance. However, this process of sequential mutation is not sufficient to explain the rapid emergence of high-level resistance; the genes for these must have been present before the drugs were introduced. In the case of trimethoprim resistance, some resistant dihydrofolate reductases came from unknown sources that are unlikely to have been bacterial in origin. We know that almost all mammalian dihydrofolate reductases are resistant to trimethoprim, hence the selective toxicity of the drug, and it is conceivable that some of these mammalian genes may have been taken up by bacteria.
Since the discovery of these bypass mechanisms in 1974, it was thought that trimethoprim and sulphonamide resistances were the only examples. In the 1990s, there was a rise in vancomycinresistant Enterococcus (VRE) strains that were becoming resistant to the final drug of choice, vancomycin. Prior to this resistance, the most common pathogenic species was Enterococcus faecalis but a greater proportion of vancomycin-resistant Enterococcus faecium emerged. The resistance was often high level and ensured the ineffectiveness of the antibiotic. Vancomycin is a natural product derived from the genus Streptomyces and it might be expected that the resistance mechanism would be destruction of the antibiotic. However, the mechanism the bond between the sH is much more complex and is, by far, the most sophisticated antibiotic-resistant mechanism found so far. The binding site of the antibiotic is the D-alanyl-D-alanine dipeptide at the end of the cross-linking pentapeptide of the peptidoglycan polymer of the cell wall. During the cross-linking process, the bond between the two alanine residues of D-alanyl-D-alanine is broken, the terminal D-alanine is removed, and the cross-link occurs with the remaining D-alanine residue. The resistance mechanism exploits this excision.
Normal synthesis of peptidoglycan requires the ligation (binding) of two D-alanine molecules to form D-alanyl-D-alanine. The resistance mechanism comprises three main genes; the first reverses this reaction so that any D-alanyl-D-alanine formed by the cell will be converted back to two molecules of D-alanine. A second enzyme converts the product of the glucose metabolism, pyruvate, into D-lactate, and a third enzyme ligates the D-lactate to D-alanine to produce D-alanyl-D-lactate. The latter is then used instead of D-alanyl-D-alanine in the formation of the pentapeptide in the synthesis of peptidoglycan. Vancomycin does not readily bind to D-alanyl-D-lactate and so the cross-linking of peptidoglycan is not inhibited. When the cross-linking occurs, the bond between the D-alanine and the D-lactate is broken, the D-lactate is removed, and the cross-linking uses the D-alanine as before. So, the composition of the final molecule is exactly the same as before; only the means of production is different.
There are a number of variations of this resistance mechanism and the genes responsible can be carried on a plasmid along with control and non-essential accessory genes. Most other plasmid encoded resistance mechanisms are a single gene product that is dominant within the cell; vancomycin resistance is different and takes over the cell’s normal synthetic machinery. Such a sophisticated mechanism of resistance could not have evolved and have been refined in the sixty years that vancomycin has been used clinically. Actually very similar mechanisms have been found in Streptomyces toyacaensis and Amycolatopsis orientalis. Despite its name vanco-, meaning ‘of unknown origin’, vancomycin was originally discovered in a strain of Amycolatopsis orientalis, an actinomycete isolated from a soil sample in Borneo. The presence of three genes, in the same orientation as those found in plasmids in clinical Enterococcus faecium, suggests that they evolved in the species producing vancomycin in order to protect it from the antibiotic it was manufacturing.
These genes were possibly captured by Streptomyces toyacaensis, another soil inhabitant, in order to protect this species against the drug. Streptomyces toyacaensis produces its own compound to attack neighbouring microorganisms but this is also a neurotoxin and too dangerous for clinical use.
Antiseptics and biocides
Most attention is directed at antibiotic resistance but actually we use a raft of other antibacterial chemicals, which we classify as biocides and subdivide as disinfectants and antiseptics. The difference between the latter two is often the concentration used: disinfectants are used at high concentrations, which can be toxic to Man, whereas antiseptics are used at much weaker concentrations as they are often applied to the extremities of the body. Should we be concerned about resistance to these compounds? The concentrations of disinfectants used clinically are often more than 1,000 times the concentration required inhibiting or killing the bacteria and they are usually indiscriminate, so will kill all microorganisms. Although resistance genes to these compounds do exist, resistance is unlikely to be a problem if the including t0R compounds are used appropriately. Even when resistance genes are present, the level of resistance they confer is far lower than concentrations that the bacteria are likely to encounter. Where problems are likely to exist is if the disinfectant is not used correctly (wrong concentration, inappropriate application, etc.). If a disinfectant is not removed completely after use and residues remain, bacteria that possess resistance genes may be able to proliferate. Similarly, if there is a large amount of organic matter, such as blood, then the efficacy of the disinfectant can be considerably weakened and resistant bacteria may proliferate; however, if the same bacteria are then placed in a solution of appropriately prepared disinfectant, then these ‘resistant’ bacteria should not survive.
The case is less clear with the antiseptics and those biocides that we incorporate into everyday objects such as cutting boards. Weaker biocides tend to be more discriminatory not just against bacteria but also the type of bacteria that they are able to inhibit. Generally speaking, they are better at controlling Gram-positive bacteria than Gram-negative. The antiseptics have more difficulty disrupting the double membrane and thicker cell wall of Gram-negative bacteria than they do with Gram-positive bacteria. The concentrations used are often, but not always, higher than those required to kill the bacteria. Resistance genes have emerged and these generally confer resistance through an efflux pump. One widely used group of antiseptics are the quaternary ammonium compounds, which are particularly effective against Gram-positive bacteria. Resistance genes, called qac encoding efflux pumps, have emerged, sometimes in multiple copies. qac genes are often found in MRSA closely linked to the genes that confer antibiotic resistance, which leads to the speculation as to whether antiseptic use can select for antibiotic resistance. This has been further compounded with the discovery that Acinetobacter baumannii, an emerging hospital pathogen, can have up to four copies of the gene. In this case they were found closely linked to genes conferring heavy-metal resistance.
The presence of the qac and other resistance genes is not always associated with decreased susceptibility to antiseptics. This may, in part, be due to the fact that we do not have clear procedures for measuring antiseptic susceptibility such as we do for antibiotics. Furthermore, antiseptic preparations often contain mixtures of compounds, including surface active agents to promote the effect of the main component. However, it may simply be due to the fact that antiseptic resistance genes are not really what we suppose them to be and are actually selected by exposure to other compounds. The close association of qac genes with heavy-metal resistance genes in Acinetobacter baumannii suggests that these genes may have been selected when the bacterium was in the environment before it became a clinical pathogen.
Clinical resistance
Microbiologists measure bacterial susceptibility by observing the ability of the antibiotic to inhibit the bacteria, when growing on agar in a Petri dish. The simplest method is to spread a bacterial culture, which has been isolated from an infection, on the surface of agar in a Petri dish. The agar contains nutrients that allow the bacteria to grow. On the bacterial ‘lawn’ are placed small circular discs made of filter paper and containing fixed and accurate amounts of antibiotics. The antibiotic diffuses out of the disc into the agar. After the Petri dish has been incubated, usually at 37°C (normal human body temperature), the bacteria will have grown over the surface of the agar except around the discs ensure that susceptibility is a fast and reasonably accurate determination of bacterial susceptibility but it is not easy to quantify. In order to achieve this, the minimum inhibitory concentration of the antibiotic has to be determined. This is usually accomplished by placing the bacteria onto agar in Petri dishes containing increasing concentrations, usually in doubling increments, of antibiotic. After incubation, the lowest concentration that is able to prevent visible growth on the agar is known as the minimum inhibitory concentration, more commonly referred to by its acronym MIC. This is a very accurate determination of bacterial susceptibility. More recently modifications of these techniques have been made and there is much research into the ability to predict susceptibility and resistance by the genes that the bacterium carries. These do not involve culture of the bacterium but rather amplification of its DNA by the polymerase chain reaction (PCR).
It is possible to identify not only specific genes but also whether they are being expressed. Depending on the combination of expressed genes in a particular species of bacteria it may be possible to ‘predict’ what the resistances would be. The advantages of these DNA-based techniques would be the speed at which they could deliver a result as there is no culture of the bacteria, and to some extent the cost, as fewer specialized personnel would be involved. The use of DNA-based techniques may seem fanciful, not least because they extrapolate in order to give a prediction; however, the culture-based techniques are also, in essence, just predictions.
So what do we mean by clinical resistance? Microbiologists are keen to report reduced susceptibility to an antibiotic or biocide but whether this can be translated in the failure of an antibiotic to cure an infection is a different matter. According to the old National Committee for Clinical Laboratory Standards (NCCLS) definition, ‘The implication of the “susceptible” category implies that an infection due to the strain may be appropriately treated with the dosage of the antimicrobial agent recommended for the type of infection and infecting species’; thus the converse is true, that a bacterium is not clinically resistant if the organism can still be controlled by clinically achievable levels of the drug.
There is a good example of this with Streptococcus pneumoniae. This bacterium can become resistant to penicillin by alterations in the target site proteins, known as the penicillin-binding proteins. Depending on the alteration, these generally can confer either a moderate or major increase in resistance to penicillin. A fully susceptible bacterium may have an MIC of 0.03mg/L, whereas a moderate increase would raise this to 1mg/L and a major increase to 16mg/L. The concentrations of penicillin reaching the lung should be in excess of 1mg/L, so bacteria with this level of ‘intermediate’ resistance may still be susceptible to treatment; indeed, this has been recommended in some cases. On the other hand, the high-level resistance should be able to overcome the clinical levels of antibiotic. Susceptibility levels of an antibiotic determined in the laboratory (in vitro) has rarely been correlated with the clinical efficacy (in vivo). There is a variety of reasons why a given antibiotic may behave differently in different patients: the heavier the patient, the less antibiotic may reach the site of infection as the antibiotic spreads through the body. No account is normally taken of patient size at prescribing. Thus the antibiotic may cure the same infection in a lighter patient but not in one who is obese. Also, infections behave slightly differently and each patient’s immune status differs. Thus we can only make broad assumptions that as a bacterium becomes more resistant as measured in vitro, it is less likely to be successfully treated in vivo.
Microbiologists continuously monitor the MICs of an antibiotic for the target bacterial populations. They usually determine the concentration of antibiotic that would inhibit 50 per cent of the population (known as the MIC50). This is the median of the sensitivity and when this figure starts to rise, it means that half the bacteria isolated are showing reductions in susceptibility. A more sensitive parameter is to determine the concentration that inhibits 90 per cent of the bacteria (known as the MIC90). This starts to rise as 10 per cent of the bacteria show reduced susceptibility. Decisions are made about the continued use of antibiotics based on information such as this. Once 10 per cent of the bacteria are showing the emergence of resistance, when previously all had been sensitive, concern might be raised. When the MRSA outbreak was at its worst in British hospitals in 1998, 40 per cent of all Staphylococcus aureus isolations were MRSA; clearly this level of resistant bacteria is far too high. It is important to note how these figures are reached; the procedure of monitoring the levels of antibiotic resistance is called surveillance. The difficulty is that the proportion of resistance will depend on the results of which bacteria are entered into the data; for example, a hospital laboratory may measure the incidences of resistance in a particular bacterial species from lung infections. These may be completely different from the resistance levels in the same species at another site of infection. In addition, the hospital laboratory may refer a proportion of its difficult bacteria to a reference laboratory; the resistance data from the reference laboratory will show a higher proportion of resistance as the population examined has already been selected because of its resistance.
The greatest discrepancies are likely to come in general practice. The majority of infections treated with antibiotics in general practice are done so empirically, that is to say without reference to sensitivity tests. If this treatment is successful the patient will not return and, in most cases, it might be assumed that the causative bacteria were susceptible. Only if the infection does not respond to treatment might a specimen be taken and then sent to a laboratory. Suppose these cases account for 10 per cent of all infections and 50 per cent of these were found to be resistant, the surveillance data from the laboratory would show that 50 per cent of the causative bacteria were antibiotic resistant whereas the true figure might be less than 10 per cent. Surveillance is not epidemiology; only if every patient with a particular infection provided a specimen could we accurately assess the true level of resistance.
Despite these minutiae, there is no doubt that clinical bacteria are becoming more resistant and therapy is threatened. During the first forty years of antibiotic use, resistance was a problem but not sufficient to force the removal of individual antibiotics from clinical use. Multi-resistance emerged as a major problem in the 1980s with MRSA and was quickly followed by other bacteria that became resistant to most antibiotics. At that time, a greater demand was being made on antibiotics; previously they had primarily been used to treat acute infections that were finally resolved by the patient’s immune system. The era of transplantation and aggressive cancer treatments causing neutropenia has meant that antibiotics have been used in patients with suppressed immune systems. The antibiotic alone has to keep the bacterial numbers low. This has allowed the emergence of bacteria, such as Acinetobacter baumannii, that have a predisposition towards resistance but can only cause infection in immune compromised patients. The reliance on antibiotics to keep undefended patients alive has placed an intolerable burden on the antibiotics that we possess. Furthermore, during the 1980s antibiotic resistance seemed to be a problem with Gram-positive bacteria, such as MRSA or VRE. This was largely because most of the antibiotics developed up to that time, for hospital use at least, were targeted at Gram-negative bacteria, which comprised the major hospital pathogens. So the failure to tackle Gram-positive bacteria allowed resistant strains to proliferate. More recently, most new antibiotics target Gram-positive bacteria and now Gram-negative bacteria are emerging as the major resistance problem.
The challenge that this causes is much greater than it was in the 1980s, as with antiseptic treatment Gram-negative bacteria are more difficult to treat with antibiotics due to their thicker cell walls. It is proving extremely difficult to find new drugs that will be sufficiently selective to control these bacteria yet be safe enough for systemic use in humans. There are significant resistance problems to the last successful anti-Gram-negative antibiotics, ciprofloxacin, the cephalosporins, and the carbapenems, especially in hospitals. Even common hospital pathogens such as Klebsiella pneumonia are becoming resistant to them all. As with MRSA, we have had to fall back on an old toxic antibiotic, colistin. Most multi-resistant Gram-negative bacteria are sensitive to colistin, but slowly resistance is beginning to develop. If it becomes firmly established, we will have almost no viable antibiotics to treat multi-resistant Gram-negative bacteria. This is likely to be a situation that is only acute in hospitals and among the immune compromised patients; however, all hospital procedures need a risk assessment and if the risk of an untreatable infection increases for certain procedures, especially through elective surgery, they will become less feasible.
Clonality of resistant bacteria
Almost all the major multi-resistant species of bacteria are not simply the sensitive bacteria that have become resistant; rather our use of antibiotics has selected out strains that have a predisposition towards resistance. MRSA is a good example of this. There is more than 20 per cent variation within the genomes of sensitive Staphylococcus aureus; the species has a core genome of around 80 per cent with differences primarily found in the variable region. When the genomes of MRSA were sequenced, it was found that this diversity was lost. So what we have been witnessing with these multi-resistant bacteria is the influence of external pressures; which, of course, have been the driving force for the evolution of all bacteria. In the case of multi-resistant bacteria in general, and MRSA in particular, the selective capability of antibiotics has provided an opportunity to observe the evolution of these bacteria.
Comparison of the genomes of all sequenced strains of MRSA has been analysed with the eBURST software. With this it is possible to recognize relationships and identify the distances between sequences of individual strains. The MRSA strains are closely related, suggesting that their common ancestry was in the less distant past. It is possible to quantify often clustered together iH when and how often certain genetic events took place. For example, the distinguishing feature of MRSA is the carriage of the mec gene that confers methicillin resistance; it appears to have migrated to MRSA strains on no more than ten occasions. However, once this gene had migrated, it was not a guarantee of success as most strains that had captured this gene have not flourished. The mec gene was first found in Sequence type (ST) 250. It was later found in the related complexes of ST8, ST239, and ST247. The complex groups ST39, ST30, and ST36 all contain MRSA; ST36 has been the most successful in the United Kingdom and is clinically known as epidemic EMRSA-16. Other features, many yet unrecognized, have been responsible for the domination of the major MRSA clones we identify today.
Similar observations have been made with the evolution of other multi-resistant bacteria, such as Streptococcus pneumoniae and Acinetobacter baumannii. In Streptococcus pneumoniae, the multiresistant strains fall within specific serogroups (originally identified by their serology but now by the genomic sequences). Specific serotypes have not only been successful because of their predisposition to resist antibiotics but also by their pathogenicity. Thirteen of the most resistant serotypes have been used to make a vaccine, Prevenar 13. The widespread use of this vaccine is expected to control these serotypes; however, there is a suggestion that new, multi-resistant serotypes are emerging to fill the void created by this vaccine, so we may see the impact of an intervention, other than an antibiotic, on the evolution of these bacteria. Similarly, although there are nearly a hundred different types of Acinetobacter baumannii, most of the strains isolated are multi-resistant versions of just four clones, known as the worldwide clones. These clones appear much more successful than the rest. The reason for this is not clear and they possess genes, currently unknown, that promote their success. Further genomic sequencing may provide this answer in the future.
Resistance in the developing world
The introduction of new resistance mechanisms into the clinical population is a relatively rare event and most of the resistance that we witness is caused by the spread of resistant bacteria. The source of new antibiotic resistance genes is a matter of hot debate. As world travel becomes quicker and easier, it is impossible to separate individual countries and it is likely that, like certain pandemic infections, hot spots occur in some countries. It is known that the incidences of resistance in bacteria from some developing countries are much higher than in the equivalent bacteria in the industrialized world. There are reasons for this: the sanitation is often poor, so there is continual infection and reinfection. Faced with this problem, many antibiotics are obtained ‘over the counter’ without reference to a health-care worker, a practice only recently outlawed in some countries of the European Union. These may either be obtained in an insubstantial number of doses or the drugs themselves may be of poor quality, only containing a small proportion of the stated doses of antibiotic. In either case, the bacteria causing the infection will often be treated with sub-inhibitory Major mechanisms of resistance used by bacteria to overcome antibiotics80R concentrations, creating a powerful selection environment for resistance development. The problem will be exacerbated as the patients are likely to be undernourished, which will compromise the immune system and make it even more difficult for the antibiotic to control the infection. There is a further issue: most of the population in developing countries lives on the land and is in close proximity to soil bacteria, a potent source of resistance genes. The mixing of these bacteria with those causing infection would allow the opportunity for the transfer of resistance genes from these soil bacteria to clinical bacteria.
Resistance in animal bacteria
There is a possibility that animals in the developing world could be a source of resistant bacteria or resistance genes; though this is not as likely as it would be with animals in the industrialized world. Most ruminants use bacteria in their gastrointestinal tract to break down the cellulose of grass and release the nutrients within. As feed for animals was modified and became richer in nutrients, the need for these bacteria decreased. They were, however, still prospering on the new feed, reducing the amount available for the host animal. In order to combat this, animals were given sub-inhibitory doses of antibiotics to suppress the growth of these commensal bacteria. These antibiotics are referred to as growth promoters and their use does significantly increase weight gain; however, the use of sub inhibitory concentrations also promotes the development of resistance. In 1969, the Swann Report outlawed the administration of antibiotics, used in human medicine, as growth promoters. In 2006, the European Union outlawed the use of all growth promoters for fear that their use was leading to antibiotic resistance. In the United States, no ban has been placed on the use of growth promoters and they are still extensively used.
Since the Second World War, many countries, including the United Kingdom, decided to massively increase their food production. This required the intensive farming of animals, especially poultry, pigs, and fish. The massive collectivization of animals leads to rapid transmission of disease from one animal to another. In order to combat this, half the antibiotics used in the United Kingdom are used to treat infections in farm animals. They are, however, used differently from clinical practice.
