Received for publication 26 November Because there was an increased incidence of sulted in a renewed emphasis on the interactions

Transcription

1 ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, May 1975, p Copyright American Society for Microbiology Vol. 7, No. 5 Printed in U.S.A. Epidemiology of Antibiotic and Heavy Metal Resistance in Bacteria: Resistance Patterns in Staphylococci Isolated from Populations in Iraq Exposed and Not Exposed to Heavy Metals or Antibiotics DAVID J. GROVES, H. SHORT, A. J. THEWAINI, AND FRANK E. YOUNG* Department of Microbiology, The University of Rochester School of Medicine and Dentistry,* Strong Memorial Hospital, Rochester, New York and Medical City Hospital, University of Baghdad, Baghdad, Iraq Received for publication 26 November 1974 Staphylococci were isolated from rural and urban populations in Iraq, which were not known to be exposed to either heavy metals or antibiotics. The antibiotic and heavy metal resistance patterns of these strains were analyzed in both mannitol-fermenting and nonfermenting strains. Over 90% of the strains were resistant to at least one of the following antibiotics: penicillin, chloramphenicol, erythromycin, tetracycline, cephalothin, lincomycin, or methicillin. In general, mannitol-fermenting strains were resistant to penicillin and cupric ions. Mannitol-negative strains were more frequently associated with mercuric ion and tetracycline resistance. Although resistance to penicillin and tetracycline can coexist, the combination of penicillin resistance and tetracycline resistance usually occurred in mannitol-negative strains. The possibility of selection of heavy metal-resistant strains due to exposure to toxic levels of methylmercury was examined. No significant increase in mercuric ion-resistant strains of staphylococci or Escherichia coli were detected in exposed populations as compared to control groups. The possible reasons for this result are discussed. The current interest in subtle effects of environmental contaminants on ecosystems has re- intercurrent infections leading to death in pa- Because there was an increased incidence of sulted in a renewed emphasis on the interactions of pollutants with microorganisms. As mata (14), it was considered particularly reletients poisoned with methylmercury at Mini- man further contaminates his own environment, he alters the milieu of those organisms for toxic amounts of heavy metals on resistance vant to determine the effects of exposure to whom he is the host. For example, selection of patterns in microorganisms. In the period from antibiotic-resistant strains in human patients September 1971 to mid-january 1972, a severe (4, 8, 19) and live stock (18) is a well recognized epidemic of methylmercury poisoning of farmers and their families occurred in Iraq, due to phenomenon. There is also evidence to indicate that there may be a correlation between the the consumption of home-made bread prepared emergence of resistance to antibiotics and heavy from seed grain treated with a methyl mercurial metals. Thus the exposure of industrial workers fungicide. A total of 6,530 cases were admitted to metallic mercury or inorganic mercury has to hospitals, and there were 459 hospital deaths been shown to be associated with an increased attributed to methylmercury poisoning (3). As colonization by antibiotic- or mercury-resistant part of an inter-university collaboration between the University of Baghdad (Iraq) and the staphylococci in the nasal passages (7). Even in populations that had not been known to have University of Rochester (U.S.A.) we explored been exposed to mercury, there is a high correlation between certain types of antibiotic resist- microbial flora of man. The results of this study the effect of methylmercury poisoning on the ance patterns and resistance to heavy metals. indicate that over 90%7o of the strains of staphylococci isolated from rural and urban populations Recent studies have demonstrated an association between resistance to penicillin, copper in Iraq, that were obtained from individuals ions, and the production of coagulase, whereas who were not known to be exposed to either resistance to mercury was more commonly heavy metals or antibiotics, were resistant to noted in strains that were resistant to tetracycline and coagulase negative (6; D. Groves and chloramphenicol, erythromycin, tetracycline, one of the following antibiotics: penicillin, F.. Young, manuscript in preparation). cephalothin, lincomycin, or methicillin. Al- 6;22

2 VOL. 7, 1975 though associations could be made between types of resistance patterns, there was no significant increase in mercuric-resistant strains of staphylococci or Escherichia coli in populations exposed to methylmercury as compared with the population selected as the control. MATERIALS AND METHODS Bacterial isolates. Strains were isolated from rectal swabs and anterior nasal swabs using Fisher Handi swabs containing modified Stuart transport medium (Fisher Scientific Co., Pittsburgh, Pa.). Outside the hospital supplies and specimens were transported in an ice chest at 4 to 10 C to minimize thermal damage. Staphylococcal strains were isolated by streaking nasal swabs on mannitol salt agar. Rectal swabs were streaked on McConkey's agar for isolation of lactose-positive enteric organisms. Antibiotic and heavy metal resistance. The assay for susceptibility and resistance to antibiotics and heavy metals were performed as described elsewhere (6; D. Groves and F. Young, manuscript in preparation). Data manipulation. Data storage, reducttn, and pattern recognition were carried out with a program designed specifically for analysis of resistance markers (6). Analysis of mercury levels. Blood samples were collected in heparinized Vacutainer tubes. The blood was analyzed for total and inorganic mercury by the atomic absorption method (9). HEAVY METAL RESISTANCE IN STAPHYLOCOCCI 623 RESULTS Antibiotic and heavy metal resistance patterns in microorganisms isolated from individuals who were not known to be exposed to antibiotics or heavy metals. To study the effect of toxicity to mercury on the plasmids in microbes which inhabit man, we attempted to obtain populations which were not known to be exposed to mercury from rural, urban, and hospital environments. The patterns of antibiotic resistance for urban and rural populations are shown in Table 1. The urban populations were derived from patients which were hospitalized or visited the out-patient clinic. None of these were known to be treated with antibiotics. The rural populations were obtained from patients visiting the popular clinic of the Mesaib al-kabir Irrigation Project (columns 5 and 6) or from rural populations in the same area (columns 7 and 8). In these samples there were very few strains that were susceptible to all antibiotics tested (35 out of 474). There was no significant difference between mannitol-positive and mannitol-negative strains. With the exception of fewer multiply resistant strains, the patterns were similar in urban and rural populations. Note that penicillin resistance occurs predominantly in mannitol-positive strains. The heavy metal resistance patterns of these staphylococcal strains are shown in Table 2. Resistance to metals is not as common as resistance to antibiotics in the same strains. Although a total of only 35 strains were susceptible to all of the antibiotics tested, 160 strains were susceptible to mercuric, cupric, and cadmium ions. These susceptible strains were largely mannitol negative. All of the strains resistant only to mercuric ions were mannitol nonfermentors, whereas strains resistant only to Cu2+ (or Cd2+) were predominantly able to ferment mannitol. Even when mercury resistance was present with resistance to other metals, the strain was usually unable to ferment mannitol. This association can be more clearly visualized by analyzing the distribution of the five predominant markers: resistance to penicillin, resistance to tetracycline, resistance to Hg2+, resistance to Cu2+, and resistance to Cd2+ independent of the others (Table 3). Because of the existence of multiply resistant strains, each TABLE 1. Distribution of mannitol fermentation and antibiotic resistance patterns Antibiotic resistance pattemsa Urban populations Rural populations Hospital Outpatients Clinic Non-clinic P C E T Cf L Dp S S S S S S S R S S S S S S S S S R S S S R S S R S S S R R S R S S S R R R R S S S Others a Abbreviations: P, penicillin; C, chloramphenicol; E, erythromycin; T, tetracycline; Cf, cephalothin; L, lincomycin; Dp, methicillin; S, susceptible; R, resistant.

3 624 GROVES ET AL. TABLE 2. Distribution of mannitol fermentation and heavy metal resistance patterns Heavy metal Urban Rural resistancea populations populations Hoptal Hosl italpatients Out- Clinic Nonclinic Hg2+ Cu2+ Cd S S S R S S S R S S S R R R S R S R S R R R R R a S, susceptible; R, resistant. TABLE 3. Analysis of independent resistance traits in mannitol fermenting and nonfermenting strains Urban Rural Independent populations populations resistancea Out- Non- Hospital patients Clinic clinic Penicillin Tetracycline Mercuric ion Cupric ion Cadmium ion athe strains are resistant to each agent independently. Thus the penicillin-resistant strains may or may not be resistant to any of the other agents tested. total population is less than the sum of the values in the column above it. The strains resistant to penicillin were distributed between mannitol-positive and mannitol-negative strains in ratios very similar to those of the total populations, as in each case the penicillin resistant strains comprise the majority of the population. In strains resistant to tetracyclines or to mercury, there is a disproportionate decrease in the fraction of mannitol-positive strains. Copper and cadmium resistance, on the other hand, causes a disproportionate increase in the mannitol-positive strains. These results (Table 3) outline the general association of resistance to penicillin and copper with mannitol fermentation as opposed to the association between resistance to tetracycline, resistance to ANTIMICROB. AGENTS CHEMOTHER. mercury, and the lack of mannitol fermentation. Therefore, the general association between resistance markers and the mannitol reaction would appear to be the same as described for other populations (D. Groves and F. E. Young, manuscript in preparation). One might expect a three-vector comparison of antibiotic resistance, metal resistance, and mannitol fermentation to show the two predominant classes to be (i) those with the mannitolpositive trait and copper as well as penicillin resistance; and (ii) those with resistance to tetracycline, resistance to mercury, and the lack of mannitol fermentation. In Table 4 we have listed, for each population, the distribution of the penicillin, tetracycline, mercury, and copper resistance patterns in fermenting and nonfermenting organisms. We have eliminated the subsets which contained less than six organisms in any pattern. In the populations resistant to only two of the test compounds (penicillin, tetracycline, mercury, and copper), the major patterns are PenRTetRHgsCus and PenRTetsCuRHgs. Thus penicillin resistance and tetracycline resistance can coexist. The classes of microorganisms with resistance only to the following pairs of agents are seldom found: penicillin and mercury; tetracycline and mercury; tetracycline and copper; or mercury and copper. Most of the strains resistant to tetracycline are also resistant to penicillin. The strains resistant only to these two antibiotics are predominantly mannitol negative. Furthermore, strains resistant to mercuric ions are usually multiply resistant strains and predominantly mannitol negative. Effect of exposure to heavy metals or antibiotics on patterns of heavy metal resistance in microorganisms. Two populations exposed to methylmercury and one population exposed to antibiotics were studied. The first population exposed to methylmercury consisted of 40 patients, largely families with children, who were kept in Medical City Hospital for treatment and research. The specimens taken from these patients were compared with control specimens taken from approximately 400 hospital patients, a subpopulation of whom had received antibiotic therapy. The second population consisted of residents of the Mesaib-al-Kabir Irrigation Project. This particular area was severely affected by methylmercury poisoning, with approximately 800 people exhibiting clinical symptoms of mercury poisoning out of a total population of 20,000. Control populations were selected from nonexposed isolated families, and from patients attending the medical clinic at Mesaib village. Both exposed and nonexposed

4 VOL. 7, 1975 TABLE 4. HEAVY METAL RESISTANCE IN STAPHYLOCOCCI 625 Major patterns of resistance to antibiotics and heavy metals Resistance patterna Urban populations Rural populations Hospital Outpatients Clinic Non-clinic P T Hg2+ Cu S S S S R S S S S S S R R R S S R S S R R R R S R R S R R R R R Others aabbreviations: Penicillin, P; tetracycline, T; susceptible, S; resistant, R. populations were validated by analysis of blood samples for total and inorganic mercury content. The resistance patterns of staphylococci to heavy metals and antibiotics were compared in populations of patients that were exposed to methylmercury and antibiotics to ascertain whether exposure influenced the types of resistance patterns in these strains. The distribution of resistance to copper, cadmium, and mercuric ions in mannitol-fermenting strains is shown in Table 5. Exposure to methylmercury did not influence the incidence of mercury-resistant strains ip the population, whereas exposure to antibiotics significantly increased the incidence of straing resistant to mercuric ions. This was most evident by an increase in strains that were multiply resistant to mercuric, copper, and cadmium ions (5 and 28% for nonexposed and antibiotic exposed, respectively). The highest incidence of mercury resistance in staphylococci occurred in the nonexposed population (Table 6). There was no significant difference in the incidence of mercury-resistant strains isolated from patients that were exposed to antibiotics or methylmercury as compared to the control population. A similar analysis of the heavy metal resistance patterns was performed on strains of staphylococci isolated from rural populations (Tables 7 and 8). The size of the nonexposed population precludes a detailed analysis. However, summation of mannitol-positive and mannitol-negative staphylococci gives zero out of 29 mercury-resistant strains from the mercury nonexposed population, compared to 12 out of 185 from the exposed population. This gives an x2 of 1.87 with d.f. = 1 and P = 0.25 to 0.1. TABLE 5. Mannitol-positive staphylococci isolated from patients in Medical City Hospital Metal resistance patterna Exposure Hg2+ Cu2+ Cd2+ None Mercury bitic S S S R S S S R S S S R R R S R S R S R R R R R a S, Susceptible; R, resistant. TABLE 6. Mannitol negative staphylococci isolated from patients in Medical City Hospital Metal resistance pattema Exposure Hg2+ Cu2+ Cd2+ None Mercury biotic S S S R S S S R S S S R R R S R S R S R R R R R , Susceptible; R, resistant. This increase is not quite significant. A larger control population might well still have had no mercury-resistant strains. Forty-five straihs with no mercury resistance in the control group

5 626 GROVES ET AL. would have made the increase in the exposed group significant with a P = 0.1. There is a possibility that the resistance of E. coli isolates might be more affected by exposure to organic mercury, as the feces contain a large portion of the eliminated mercury, thus exposing fecal microorganisms to elevated levels. Table 9 shows the data for mercury resistance and susceptibility of the E. coli isolated from fecal samples and rectal swabs of both exposed and nonexposed populations. As observed with staphylococci there was a higher incidence of TABLE 7. Heavy metal resistance of mannitol-positive staphylococci isolated from rural subjects Metal resistance patterna Hg2+ CU2+ Cd2+ S S S R S S S R S S S R R R S R S R S R R R R R a R, Resistant; S, susceptible. TABLE 8. Heavy metal resistance of mannitol-negative staphylococci isolated from rural subjects Metal resistance patterna Exposure Hg2+ Cu2+ Cd2+ None Mercury S S S 8 28 R S S 0 1 S R S 3 54 S S R 0 0 R R S 0 6 R S R 0 0 S R R 0 0 R R R 0 2 ar, Resistant; S, susceptible. TABLE 9. Effect of exposure to methylmercury on resistance of mercuric ions in strains of Escherichia coli Non-clinic Hospital (no Resistance to Mesaib project antibiotic mercuric ion exposure) Rb Sb a Exposure to mercury indicated by + and -. b R, Resistant; S, susceptible. ANTIMICROB. AGENTS CHEMOTrHER. resistance to mercury in the organisms isolated from individuals in the urban environment. Even with the small sample studied it is unlikely that poisoning with methylmercury shifted significantly the levels of resistance to mercuric ions at the time of sampling of the population. DISCUSSION Microorganisms frequently contain extrachromosomal elements that determine resistance to toxic compounds such as antibiotics and heavy metals, or the determinants for genes that regulate complex pathways of biodegradation such as the metabolism of camphor in Pseudomonas putida (15). In staphylococci a single genetic determinant, coagulase, has been utilized traditionally to separate Staphylococcus aureus from Staphylococcus epidermidis. Other traits such as mannitol fermentation (13), phage sensitivity, and biotyping (2) have reinforced this grouping. The observation that genetic markers can be exchanged between the S. aureus and S. epidermidis (20) indicates that this grouping may be quite artificial. Nevertheless, we utilized mannitol fermentation as an additional trait in the analysis of the distribution of heavy metal resistance among staphylococci. If the possession of resistance to mercury produced a selective advantage to microorganisms inhabiting poisoned individuals, it would not be unreasonable to expect an increased incidence of mercury-resistant strains either in the staphylococci which usually inhabit the nasopharynx (S. aureus) or those which usually inhabit the skin (S. epidermidis). However, no significant increase in the incidence of mercuryresistant strains occurred in individuals that were severely poisoned with mercury. A number of explanations should be considered in evaluating this result. First, the exposure to mercury was limited in duration. Within 2 months of the onset of ingestion of contaminated bread, symptoms appeared and the treated grain was recovered. Thus the blood levels rose to a maximum, then decayed with a half-life of approximately 70 days (3). Second, there was an interval of at least 6 months between the consumption of the contaminated bread and the sampling of the population for analysis of the microbial flora. Nevertheless, the total mercury levels in the blood of exposed subjects were still elevated, with a range of 100 ng/ml to 1,350 ng/ml. Even when a limited population was studied in which the blood levels were greater than 500 ng/ml, there was no significant increase in the mercury-resistant fraction of E.

6 VOL. 7, 1975 coli or staphylococci. Third, the blood level values used are total mercury levels, with the inorganic mercury composing less than 10% of this value (3). Fourth, although strains of staphylococci which are resistant to mercury ions are also resistant to phenyl mercuric ions, there is no relationship to their resistance to methylmercury (G. D. Groves and F. E. Young, manuscript in preparation). Thus, poisoning with alkyl mercurial compounds may not provide selective pressure for resistance to mercuric ions even though the elevated circulating inorganic mercuric levels in the blood show that significant degradation to inorganic mercury occurs. Finally, the high frequency of antibiotic-resistant staphylococci, even among those not known to have been treated with antibiotics, would obscure subtle changes in the microbial flora. For instance, antibiotic-resistant staphylococci occurred with a frequency of greater than 0.90 with an incidence of 0.86 for penicillin resistance. The only trend that indicated a selective effect of mercury was in the rural population; however, the low numbers of isolates precluded a detailed analysis (Tables 7 and 8). In addition, the infrequent occurrence of mercuryresistant organisms in the rural environment where the original exposure to organo-mercurials occurred would result in a limited gene pool from which mercury-resistant organisms might be selected. The ideal population for study would be carriers of mercury-susceptible staphylococci and E. coli who become exposed to high levels of inorganic or organic mercury with sampling prior to, during, and after exposure. It is also necessary to have a pool of mercuryresistant strains in the environment for selection by the elevated mercury levels. Such ideal human populations are not possible, but model animal systems are under consideration. Studies are also under way to sample populations with chronic exposure to organo-mercurials, as well as chronic and transitory exposure to inorganic mercury. Many of the genes which regulate antibiotic and heavy metal resistance in staphylococci reside on plasmids. Although there appears to be no incompatibility between plasmid-borne resistance to penicillin, tetracycline, chloramphenicol, and erythromycin in S. aureus (5, 17), little is known about the interactions between markers that reside on different plasmids and HEAVY METAL RESISTANCE IN STAPHYLOCOCCI 627 how these plasmids interact to form multiple resistant strains. The most extensively studied plasmids in S. aureus are the penicillinase-containing plasmids of phage groups I and II. Resistance to heavy metals such as mercuric and cadmium ions are usually linked to penicillinase on these plasmids (11, 12, 16). Tetracycline resistance can be determined by a plasmid which is lost or transduced independently of the penicillinase-containing plasmids (1, 10). The data in this study demonstrate that patterns of antibiotic and heavy metal resistance vary between coagulase-positive or mannitol-fermenting and coagulase-negative or mannitol-noniermenting strains. For instance, resistance to penicillin and copper ions are most frequently associated with the presence of mannitol fermentation, whereas the resistance to tetracycline and mercuric ions is most frequently found in mannitol nonfermenting strains. Although it is possible to have resistance to penicillin, tetracycline, mercuric ions and copper ions in the same strain, this most frequently occurs when the organism cannot ferment mannitol. It may be of interest to note that similar association between the resistance to mercury and tetracycline in coagulase-negative strains in contrast to the resistance to penicillin and copper in coagulase-positive strains was observed in the microbiology laboratory at the University of Rochester. Detailed studies are in progress to determine whether the insights from the epidemiological survey may aid in obtaining a more detailed understanding of the incompatibility and interaction of plasmids in pathogenic organisms. ACKNOWLEDGMENTS We are grateful to the Scientific Committee on Mercury Poisoning in the Baghdad Medical College for allowing us to publish these results. The analytical data on mercury were supplied by the Mercury Research Laboratory under the direction of Hashim I. Dhahir. We wish to acknowledge P. Dhahir and M. R. Greenwood for supervising the analytical team and for data processing. The analytical determinations were made by Ilham M. Al-Jubouri, Amir Khayat, Selma M. Matook, and Mansour Al-Muntasir. We wish to also extend our thanks to Ammar Abdul Razzak and Muhammed Ali Majeed for their invaluable help in collecting the samples and in other aspects of the work. We also wish to express our appreciation to Buthaina Al-Nakash, Agnes Haik, Suhaila Saadalla, Rasoul Al-Dabagh, Hashim Al-Mosawi, and other workers in the Microbiology Laboratory in the Medical City Hospital for their indispensable help. David L. Hollis contributed to the technical aspects of the study at the University of Rochester. This research was supported in part by National Science Foundation (RANN) grant GI and in part by funds supplied by Baghdad University. We are grateful to Bioquest Division of Baltimore Biological Laboratory and Difco Laboratories for their kind donation of supplies and media. LITERATURE CITED 1. Asheshov, E. H Chromosomal location of the genetic elements controlling penicillinase production in a strain of Staphylococcus aureus. Ntature (London) 210: Baird-Parker, A. C Staphylococci and their classification. Ann. N. Y. Acad. Sci. 128: Bakir, F., S. F. Damluji, L. Amin-zaki, M. Murtadha, A.

7 628 GROVES ET AL. Khalidi, N. Y. Al-Rawi, S. Tikriti, H. I. Dhahir, T. W. Clarkson, J. C. Smith, and R. A. Doherty Methylmercury poisoning in Iraq: an interuniversity report. Science 181: Bentley, D. W., J. J. Hahn, and M. H. Lepper Transmission of chloramphenicol-resistant Staphylococcus epidermidis: epidemiologic and laboratory studies. J. Infect. Dis. 122: Dyke, K. G. H., and M. H. Richmond Occurrence of various types of penicillinase plasmids among hospital staphylococci. J. Clin. Pathol. 20: Groves, D. J., L. Maroglio, C. W. Merriam, and F. E. Young Epidemiology of antibiotic and heavy metal resistance in bacteria: a computer-based data system. Comput. Programs Biomed. 3: Hall, B. M Mercury resistance of Staphylococcus aureus. J. Hyg. 68: Lepper, M. H., P. Tillman, and R. Devetsky Patterns of transmission of staphylococci. Ann. N. Y. Acad. Sci. 128: Magos, L., and T. W. Clarkson Atomic-absorption determination of total, inorganic, and organic mercury in blood. J. Assoc. Off. Anal. Chem. 55: May, J. W., R. H. Houghton, and C. J. Perret The effect of growth at elevated temperatures on some heritable properties of Staphylococcus aureus. J. Gen. Microbiol. 37: Moore, B A new screen test and selective medium for the rapid detection of epidemic strains of Staphylococcus aureus. Lancet 2: Novick, R. P., and C. Roth Plasmid-linked resistance to inorganic salts in Staphylococcus aureus. J. ANTIMICROB. AGENTS CHEMOTHER. Bacteriol. 95: Raymond, E. A., and W. H. Traub Identification of staphylococci isolated from clinical material. Appl. Microbiol. 19: Report from an expert group Methylmercury in fish. A toxicologic-epidemiologic evaluation of risks, p Nordisk Hygienisk Tidskrift Supplement 4, Stockholm. 15. Rheinwald, J. G., A. M. Chakrabarty, and I. C. Gunsalus A transmissible plasmid controlling camphor oxidation in Pseudomonas putida. Proc. Natl. Acad. Sci. U.S.A. 70: Richmond, M. H., and M. John Cotransduction by a staphylococcal phage of the genes responsible for penicillinase synthesis and resistance to mercury salts. Nature (London) 202: Richmond, M. H., M. T. Parker, M. P. Jevons, and M. John High penicillinase production correlated with multiple antibiotic resistance in Staphylococcus aureus. Lancet 1: Smith, W. H Veterinary implications of transfer activity, p In G. E. W. Wolstenholme and M. O'Conner (ed.), Ciba foundation symposium on bacterial episomes and plasmids. Little, Brown and Co., Boston. 19. Watanabe, T Infective heredity of multiple drug resistance. Bacteriol. Rev. 27: Yu, L., and J. N. Baldwin Intraspecific transduction in Staphylococcus epidermidis and interspecific transduction between Staphylococcus aureus and Staphylococcus epidermidis. Can. J. Microbiol. 17: Downloaded from on January 14, 2019 by guest