RAPID IDENTIFICATION OF CHOLERA VIBRIOS WITH FLUORESCENT ANTIBODY

Transcription

1 RAPID IDENTIFICATION OF CHOLERA VIBRIOS WITH FLUORESCENT ANTIBODY RICHARD A. FINKELSTEIN AND EUGENE H. LABREC Division of Immunology, Walter Reed Army Institute of Research, Washington, D. C. Received for publication June 15, 1959 The fluorescent antibody technique has been used experimentally in a number of laboratories for the detection of a variety of bacterial and viral agents (Thomason et al., 1956; Moody et al., 1958; Whitaker et al., 1958; Liu, 1956). Recently, an attempt was made by Thomason et al. (1958) to apply the technique directly in the detection of enteric pathogens, including Salmonella species and Vibrio comma (Vibrio cholerae), in stool specimens, but the occurrence of extrageneric cross reactions made the results equivocal. Similar experiences were noted by LaBrec et al. (1958) with Shigellae when fluorescein-labeled grouping sera were employed although satisfactory specificity was obtained by the use of labeled, adsorbed, type-specific antisera. The present investigation was initiated to determine if the fluorescent antibody technique might facilitate early recognition and identification of Vibrio comma in human stools. It was considered that the problem of extrageneric cross reactions might be circumvented by exploiting an unusual physiological attribute of the cholera vibrio: its ability to grow in media of high alkalinity, to supplement the fluorescent antibody technique with selective enrichment. Our observations in this regard provide the basis for the present report. MATERIALS AND METHODS Cultures. Most of the cultures used in this investigation were obtained from the culture collection of the Walter Reed Army Institute of Research (WRAIR). The Vibrio strains used were originally isolated in India, Thailand, Egypt, China, and Japan and included known Ogawa and Inaba serotypes as well as El Tor vibrios and "intermediate" forms. A series of "rough" (R) and "rho" (p) strains was received from Dr. C. E. Lankford (Lankford, 1958). Antisera. Rabbit hyperimmune V. comma antisera were prepared essentially according to the technique of Burrows and Pollitzer (1958) 886 using steamed (2 psi per 2 hr) antigens of the National Institutes of Health (NIH) reference strains NIH 35 (Inaba) and NIH 41 (Ogawa) (WRAIR strains 20-A-10 and 20-A-11, respectively). The homologous antigens were agglutinated by these antisera at dilutions of 1:2560 to 1:10,240. Additional unadsorbed Ogawa, Inaba, and rough antisera were obtained from Dr. Sanders T. Lyles. Cholera grouping sera were prepared by mixing equal volumes of unadsorbed Inaba and Ogawa antisera. Typespecific Inaba and Ogawa antisera were prepared by adsorbing the antisera with the heterologous acetone-killed and dried cells. In our experience this did not significantly affect the homologous agglutinin titer. Fluorescein-labelino of sera. Both grouping and type-specific antisera were conjugated with fluorescein isothiocyanate using the procedure described by Marshall et al. (1958). Following conjugation, the free dye was removed by dialysis against 0.01 M phosphate buffered saline, ph 7.0 to 7.2. On occasion, the free and adsorbed dye was removed from the fluorescein-labeled antisera by shaking with charcoal (0.05 g per ml) for 1 hr (Chadwick et al., 1958) prior to dialysis against buffered saline. Finally, the conjugated sera were sterilized by filtration through Millipore filters. Some additional unadsorbed fluorescein-labeled sera against Inaba, Ogawa, and El Tor vibrios were kindly supplied by Captain J. Marshall. Bacterial smears. Bacterial smears for immunocytochemical staining were prepared from cultures grown on a variety of media; meat extract agar, brain heart infusion agar, 2 per cent glycerol- 1 per cent glucose-peptolie-beef extract agar ("2/1 agar," Lyles and Gardner, 1958), brain heart infusion broth, alkaline peptone broth, and the simple defined medium of Finkelstein and Lankford (1955). The medium used did not appear to affect the subsequent staining reactions with the fluorescein-labeled antisera.

2 1959] IDENTIFICATION OF CHOLERA VIBRIOS 887 Heat fixed smears were stained with conjugated antiserum for 20 min in a humid atmosphere and washed thoroughly by agitation for 15 min with two changes of buffered saline at ph 7.0 to 7.2. After cover slips were mounted with buffered glycerol, the smears were examined with a Leitz Ortholux microscope fitted with a darkfield condenser and appropriate filters. The light source was a Phillips 150 w high pressure mercury arc burner. To simulate cholera infected human stools, 10 per cent suspensions in saline were made of stool specimens submitted to Walter Reed Army Hospital clinical laboratory and from healthy volunteers. To 0.9-ml portions of the lightly centrifuged stool suspensions were added 0.1 ml samples of suspensions of cholera vibrios diluted to contain the desired concentrations of organisms. Direct smears were made from these by depositing droplets on slides by means of a standard 3-mm (26 gauge) platinum loop which delivered approximately ml in a circumscribed area. To facilitate locating the smears for microscopic examination, the desired number of circles of approximately 3-mm diameter were scratched on the slide with a diamond point scriber. In this manner, a large number of relatively standard droplet smears could be made on a single slide. The droplets were dried in air, heat-fixed, and stained. Smears were also made of the feces, cecal contents, and ileal contents of normal guinea pigs and guinea pigs experimentally infected with cholera by the technique described by Formal et al. (1958) with Shigella. Selective enrichment. Selective enrichment of cholera vibrios was achieved by inoculating alkaline peptone broth (1 per cent peptone, 0.5 per cent NaCl, ph 8.5 to 9.0), 9.0 ml, with 0.1 ml of the infected stool suspensions described above. At intervals during incubation at 37 C, smears were made with the standard loop from the surface growth in the peptone water. Simultaneously, the growth was streaked on 2/1 agar; the colony morphology of V. comma on this agar, when observed in oblique transmitted light with a stereoscopic microscope, is characteristic and distinct from other enteric bacteria which may grow on the medium (Lankford, 1953, unpublished data). On occasion, the identification of colonies as V. comma was verified by slide agglutination tests. RESULTS Preliminary observations. Initial observations revealed that cholera vibrios in pure culture stained with fluorescein conjugated immune rabbit cholera grouping sera exhibit the characteristic specific bright yellow-green fluorescence. Monovalent conjugated sera clearly distinguished Inaba and Ogawa serotypes. Fluorescence of vibrios could not be detected following treatment with two different preparations of fluoresceinlabeled normal rabbit serum. To demonstrate further the specificity of the fluorescent staining, an attempt was made to block the reaction by prior treatment of the fixed smears with unlabeled immune serum. Presumably, this should saturate the free antigenic sites and eliminate subsequent attachment of conjugated antibody. However, smears prepared in this manner exhibited relatively undiminished fluorescence. Goldman (1957), reported similar difficulty using this "two-step" inhibition in proving specificity of fluorescent antisera directed against Toxoplasma. If, instead of the two step technique, the smears are treated with a mixture of undiluted immune serum and fluorescein-labeled antibody diluted to the point where further dilution would result in diminishing brightness, specific fluorescence of the vibrios is prevented. Controls stained with the equivalent final dilution of conjugated antiserum approximate the brightness of cells stained with undiluted conjugated antiserum. To determine the ability of fluorescent antibody to stain cholera vibrios in biological material, smears were prepared from the intestinal contents of moribund guinea pigs experimentally infected with cholera. After treatment with conjugated sera, specifically stained cholera vibrios in large numbers were readily recognized; paraliel culture on 2/1 agar disclosed that 90 per cent or more of the colonies were V. comma. However, in smears of suspensions of feces from healthy animals, microorganisms of various forms stained with varying degrees of brightness. There was no difficulty in recognizing the cholera infected material because of the large number of typical brightly fluorescing vibrios. However, it was considered that there was sufficient nonspecific staining to render interpretations on minimally contaminated specimens equivocal. Adsorption of the conjugated sera with bone marrow powder and guinea pig tissue powder did not significantly

3 888 FINKELSTEIN AND LABREC [VOL. 78 affect the degree of nonspecific staining. Adsorption with acetone precipitated human feces reduced, but did not eliminate, the nonspecific reactions. Observations on human stools artificially contaminated with cholera vibrios. Since the objective of this investigation was the prompt recognition of cholera vibrios, possibly in small numbers, in human stools, an attempt was made to evaluate the sensitivity of the technique. Samples of 10 per cent human fecal suspension were inoculated with 10-fold serial dilutions of a suspension of V. comma so that final portions contained 0 to 109 vibrios per ml. These suspensions were randomized, coded, applied to slides with the standard loop, stained, and examined, independently, by two observers. When the suspensions contained 107 or more vibrios per ml, there was no difficulty in classifying the smears as positive by the fluorescent technique. Below this level of contamination, however, the results were equivocal. Occasionally, negative samples were reported as doubtfully positive (few organisms resembling vibrios seen) and smears from suspensions with low levels of contamination were reported as negative. To enhance the sensitivity and reduce the nonspecificity of the fluorescent technique, the principle of enrichment culture was applied. Nine-tenth-ml samples of 20 human stool suspensions were inoculated with 107 V. comma in 0.1 ml volume. Smears were made directly from the contaminated and control specimens, and simultaneously, alkaline peptone broth (ph 9.0, 9-ml) tubes were seeded with 0.1 ml of each specimen. After 4 and 10 hr incubation, droplet smears were made from the surface growth in the broth cultures and 2/1 agar plates were streaked. On examination of the direct smears, many of the control uninoculated stools could falsely be classified as positive because of the presence of significant amounts of fluorescent material which could be confused with specifically stained V. comma. On the other hand, all of the direct smears of stools contaminated with 107 V. comnmna were distinctly positive. After 4 and 10 hr incubation, stained smears from all of the peptone broth cultures of the inoculated specimens contained tremendous numbers of brightly fluorescing vibrios in every microscopic field. Approximately 75 to 100 per cent of the colonies on 2/1 agar plates streaked from each of the inoculated specimens were V. comma. In stained droplet smears from the peptone broth cultures of the uninoculated stools, no organisms which could be confused with cholera vibrios could be detected. In three of these cultures, fluorescent organisms were detected, but in each of these, in marked contrast to the V. comma infected material, the fluoresence was weak and could only be seen with difficulty. In one case, the organisms were large ovoid cocci and in the other two, long thin rods. To determine the sensitivity of the combined technique, suspensions of five human stools were inoculated with V. comma Ogawa (strain 20-A-1) so that they contained from 2.5 X 101 to 2.5 X 106 cholera vibrios per ml. Alkaline peptone broth tubes (ph 8.5) were inoculated with 0.1 ml of each sample and streaked to 2/1 agar after 3, 6, and 9 hr incubation. Droplet smears for fluorescent antibody staining were made at the same times. From the results summarized in table 1, it can be seen that V. comma can be detected with confidence by the combined enrichment-fluorescent antibody technique within 9 hr in stool suspensions initially seeded with 2.5 X 102 or more vibrios per ml. When the starting population of V. comma was 2.5 x 101, vibrios were definitely identified in 3 out of 5 stools. With only 6 hr incubation in peptone broth, the lowest limit for reliable diagnosis was about 2.5 X 101 organisms. By cultural methods alone, the vibrios may be recovered earlier in the incubation period of the peptone broth enrichment, but an additional incubation time of 12 hr or more is needed for recognizable colonies to develop on 2/1 agar or other plating media. No instance of a false-positive staining reaction has been observed in the human stools examined with this two-step technique. In addition, specific type identification, i.e., in this case, bright fluorescence in Ogawa serum and nonreactivity with conjugated Inaba serum, was obtained by using duplicate smears. It should be noted that in certain stool suspensions, there was a tendency for the proportion of vibrio colonies recovered to decrease during the incubation of the peptone broth enrichment cultures. In some cases, the growth of vibrios in the peptone broth appeared to have a synergistic effect on growth of other organisms since no colonies were recovered on 2/1 agar from the peptone broth inoculated with the control stool

4 1959] IDENTIFICATION OF CHOLERA VIBRIOS 889 TABLE 1 Selective enrichment and fluorescent antibody staining in the rapid detection of Vibrio comma in artificially contaminated human stool suspensions Hr Incubation in Peptone Results Initial V. comma Population* Broth X X X 10' 2.5 X 10' 2.5 X X 106 Positive 0/Ot 0/0 0/1 0/4 0/4 0/4 1/5 3 Doubtful 0/0 0/0 0/0 0/0 0/0 3/0 4/0 Negative 5/5 5/5 5/4 5/1 5/1 2/1 0/0 Positive 0/0 0/3 1/5 4/5 5/5 5/5 5/5 6 Doubtful 0/0 0/0 2/0 1/0 0/0 0/0 0/0 Negative 5/5 5/2 2/0 0/0 0/0 0/0 0/0 Positive 0/0 3/5 4/5 5/5 5/5 5/5 5/5 9 Doubtful 0/0 1/0 1/0 0/0 0/0 0/0 0/0 Negative 5/5 1/0 0/0 0/0 0/0 0/0 0/0 * Per ml 10 per cent stool suspension. t Numerator: V. comma identified by fluorescent antibody staining; x number of times in 5 specimens. Denominator: V. comma identified by colony morphology on 2/1 agar; x number of times in 5 specimens. suspensions, whereas some colonies in addition to V. comma were isolated from the counterparts seeded with V. comma. In other cases, cholera vibrios appeared to multiply at the expense of other organisms present since extraneous organisms were recovered from control preparations but not from the parallel cultures inoculated with vibrios. However, no colonies morphologically resembling V. comma were observed when control specimens were plated. Presence of the fecal material however, did not seem to affect the recovery of the V. comma by the fluorescent antibody or the cultural technique with the short incubation periods used, since essentially the same degree of sensitivity was noted when parallel peptone broth cultures of vibrios without feces were studied by the same techniques. Examination of a number of strains of cholera vibrios, related forms, and other enteric organisms. Smears were made in triplicate of a number of strains of V. comma and other enteric bacteria, all of which were randomized and coded. One smear of each group was stained with cholera grouping serum, the other two, with specific Inaba and Ogawa sera, respectively. All of the 23 cholera strains tested were detected with grouping serum. Of these, 11 serologically typical Ogawa strains and 9 typical Inaba strains were identified with labeled monovalent typing sera. In three instances, fluorescent staining reactions with monovalent sera were equivocal; two of these organisms failed to react typically in slide agglutination tests and fluoresced following staining with conjugated rough antiserum which did not stain the reference strains, 20-A-10 and 20-A-11. Therefore, partial serological degradation of the strains may explain in part the indeterminate results of the fluorescent staining. Three stains, classified as "intermediate" or "paracholera" vibrios when isolated, were stained by grouping serum but not typing serum; i. e., they reacted as though they contained the O group I antigen (A) but were deficient in typespecific (B and C) antigens (Pollitzer and Burrows, 1955). Only one of these strains fixed labeled rough antibody. Additional investigation would be required to define the antigenic nature of these aberrant strains. R and p strains of V. comma exhibited varying degrees of fluorescence following staining with conjugated unadsorbed grouping serum. This activity was markedly enhanced if the smears were prepared from cells previously steamed as in the preparation of immunizing antigens. This observation is in accord with the recent findings of Lankford (1958) and Lyles and Gardner (1958) that smooth V. comma antigens may stimulate the production of significant amounts of rough and rho antibodies, the recognition of which is facilitated by the use of steamed agglutinating antigens. This

5 890 FINKELSTEIN AND LABREC [vol. 78 activity is removed by adsorption with steamed R and p antigens without detectable loss of the staining activity of the serum for smooth strains. El Tor vibrios, which are capable of causing epidemic enteritis, are detected by cholera grouping sera and generally appear to stain according to their relationship to the true cholera serotypes. True cholera vibrios are likewise stained by fluorescein-labeled unadsorbed El Tor antisera. Representative strains of other enteric bacteria including Vibrio proteus, Proteus OX2, Proteus morganii, Escherichia coli, Salmonella typhosa (Vi and 0), Salmonella enteritidis, and Shigellaflexneri were not stained either by cholera grouping serum or type-specific cholera antisera. DISCUSSION Conventional culture techniques for the isolation of cholera vibrios require, at the minimum, an overnight incubation period before the identification of an isolated colony as V. comma may begin (Burrows and Pollitzer, 1958). The results of the present investigation indicate that the selective enrichment-fluorescent antibody technique may provide a rapid, reliable, and sensitive means of detecting cholera vibrios in human stool suspensions: the time required to establish a diagnosis is reduced to a total of from less than 6 hr to 9 hr, depending on the initial population of cholera vibrios in the infected specimen, and large numbers of specimens may be handled with facility. The difficulty in establishing a typespecific diagnosis of occasional strains by the fluorescent antibody technique is likewise encountered using conventional serological methods. It should be recognized that the procedure has only been applied, thus far, to simulated pathological material and the value of these observations can only be assessed by field trial on specimens from naturally infected cases and carriers. It is believed that this technique will facilitate screening for carriers during inter-epidemic periods and be useful in locating carriers and diagnosing cases during epidemic waves. The ability to diagnose cases more promptly in the early stages of an epidemic outbreak would be of immeasurable value in enabling proper public health measures to be initiated at the earliest possible moment. The ability of labeled unadsorbed grouping serum to stain rough strains of V. comma poses a question which is difficult to answer in the light of present knowledge. Whether this activity should be left in conjugated sera used for screening for cholera cases and carriers would depend on the desired result. On the one hand, it may be advisable to remove the rough staining activity to increase the specificity of the recognition of smooth, presumably virulent, strains. On the other hand, there is some evidence that rough (avirulent?) strains of V. comma may give rise to smooth (virulent?) variants and the detection of rough strains may therefore assume public health significance. The problem may be resolved for the present by the use of parallel smears, one stained with grouping serum from which rough antibody has been removed and one stained with conjugated rough antiserum to detect the possibly significant rough variants which may be confirmed by conventional bacteriological techniques (Lankford, 1958). ACKNOWLEDGMENTS The authors are indebted to Lt. Colonel A. S. Benenson, MC, and Dr. Samuel B. Formal, Walter Reed Army Institute of Research, for helpful suggestions and criticism. SUMMARY By combining selective enrichment and fluorescent antibody techniques, a method has been devised which facilitates the rapid recognition and serological identification of small numbers of cholera vibrios (Vibrio comma) in human stool suspensions. This procedure minimizes the nonspecific reactions encountered in the direct examination of stool specimens. ADDENDUM After this manuscript was drafted, Thomason, Cherry, and Edwards (J. Bacteriol., 77, , 1959) reported results of fluorescent antibody staining of stool smears from typhoid carriers and normal persons. Using conjugated cholera 0 antiserum, they found fluorescent organisms or objects in 17 of 25 specimens from typhoid carriers and 21 out of 25 normal stool smears. One lot of fluorescein labeled cholera antiserum stained 5 cultures (3 strains of E. coli, 1 Proteus, and 1 E. freundii isolated from normal stool) of the large series of pure cultures tested; conjugates of other V. comma antisera failed to stain these organisms. These observations serve to illustrate

6 1959] IDENTIFICATION OF CHOLERA VIBRIOS 891 the need for augmentation or enforcement of the fluorescent antibody technique to achieve greater specificity in diagnosing enteric pathogens. REFERENCES BURROWS, W. AND POLLITZER, R Laboratory diagnosis of cholera. Bull. World Health Organization, 18, CHADWICK, C. S., McENTEGART, M. G., AND NAIRN, R. C Fluorescent protein tracers. A simple alternative to fluorescein. Lancet, 1, FINKELSTEIN, R. A. AND LANKFORD, C. E Nutrient requirements of Vibrio cholerae. Bacteriol. Proc., 1955, 49. FORMAL, S. B., DAMMIN, G. J., LABREC, E. H., AND SCHNEIDER, H Experimental Shigella infections: Characteristics of a fatal infection produced in guinea pigs. J. Bacteriol., 75, GOLDMAN, M Staining Toxoplasma gondii with fluorescein-labelled antibody. II. A new serological test for antibodies to Toxoplasma based on inhibition of specific staining. J. Exptl. Med., 105, LABREC, E. H., FORMAL, S. B., AND SCHNEIDER, H Serological analysis of Shigella flexneri by the fluorescent antibody technique. Bacteriol. Proc., 1958, 135. LANKFORD, C. E The serologic identification of Vibrio cholerae. J. Microbiol. Soc. Thailand, 2, Liu, C Rapid diagnosis of human influenza infection from nasal smears by means of fluorescein-labeled antibody. Proc. Soc. Exptl. Biol. Med., 92, LYLES, S. T. AND GARDNER, E. W Studies on the antigenic structure of Vibrio cholerae. 1. The effects of heat on the agglutinating antigens. J. Infectious Diseases, 103, MARSHALL, J. D., EVELAND, W. C., AND SMITH, C.- W Superiority of fluorescein isothiocyanate (Riggs) for fluorescent-antibody technic with modification of its application. Proc. Soc. Exptl. Biol. Med., 98, MOODY, M. D., ELLIS, E. C., AND UPDYKE, E. L Staining bacterial smears with fluorescent antibody. IV. Grouping streptococci with fluorescent antibody. J. Bacteriol., 75, POLLITZER, R. AND BURROWS, W Cholera studies. 4. Problems in Immunology. Bull. World Health Organization, 12, THOMASON, B. M., CHERRY, W. B., AND EDWARDS, P. R Fluorescent antibody examination of fecal smears for enteric pathogens. Bacteriol. Proc., 1958, 134. THOMASON, B. H., MOODY, M. D., AND GOLDMAN, M Staining bacterial smears with fluorescent antibody. II. Rapid detection of varying numbers of Malleomyces pseudomallei in contaminated material and infected animals. J. Bacteriol., 72, WHITAKER, J., PAGE, R. H., STULBERG, C. S., AND ZUELZER, W. W Rapid identification of enteropathogenic Escherichia coli 127:B8 by the fluorescent antibody technique. A. M. A. J. Diseases Children, 95, 1-8.