Comparison of conventional and immunofluorescent techniques for the detection of Entamoeba histolytica in rectal biopsies

GASTROENTEROLOGY

78:435-439,

1980

Comparison of Conventional and Immunofluorescent Techniques for the Detection of Entamoeba histolyfica in Rectal Biopsies R. GILMAN, M. ISLAM, S. PASCHI, and F. AHMAD

J. GOLEBURN,

The Johns Hopkins University International Center for Medical Research, Baltimore, Maryland; The Department of Pathology, Dacca Medical College, Dacca, Bangladesh: and The International Centre for Diarrhoeal Disease Research, Bangladesh, Dacca, Bangladesh.

Two comparative studies of six staining techniques in detection of Entamoeba histolytica in rectal biopsies were performed. The techniques evaluated were direct and indirect fluorescent antibody technique, and four conventional stains: PAS, H 6 E, Trichrome, and PTAH). On ethanol-fixed tissue, the direct fluorescent antibody technique was the most sensitive, detecting amoebas in 84% of rectal biopsies from 19 aspirate-positive dysentery patients. PAS staining was specific, almost as sensitive, and amoebas could be rapidly detected. The other three conventional methods, however, detected amoebas in only 53-58% of the aspirate-positive patients and required considerable efforts to find the amoebas. The indirect fluorescent antibody technique performed on ethanol-fixed tissues was comparable to the direct technique. The same technique used on formalin-fixed tissues had high background staining, but amoebas could still be detected more efficiently than with conventional staining. Following 20 hr or more treatment with metronidazole, no patient had amoebas detectable in their rectal biopsies by any technique used. Patients with amoebic dysentery and necrotic ulcers on procto-

Received August 8, 1979. Accepted October 22, 1979. Address requests for reprints to: Dr. Robert Gilman, The Johns Hopkins University ICMR, c/o Baltimore City Hospital, 4940 Eastern Avenue, Baltimore, Maryland 21224. This study was supported by a grant from the National Institutes of Health 5 R07 AI10048-18. The authors would like to thank the staff of the International Centre for Diarrhoeal Disease Research, Bangladesh Hospital for their cooperation, to Mr. G. Modal and M. Wahid for technical assistance, and Drs. R. B. Sack, F. Koster, and J. Yardley for advice. 0 1980 by the American Gastroenterological Association 0016-5085/80/030435-05$02.25

scopy had amoebas detected more frequently in rectal biopsy than did patients with mild disease as determined by proctoscopy. Entamoeba histolytica in rectal biopsies can be detected most easily in ethanol-fixed tissue stained with either a fluorescent antibody or with PAS. Jf amoebas are not detected in the rectal biopsy, amoebic colitis cannot be excluded. We believe that the examination of stool or anoscopic aspirate for hematophagous motile trophozoites of Entamoeba histolytica is the simplest, easiest, and quickest method of establishing the diagnosis of amoebic dysentery. However, the diagnosis of amoebic colitis is often overlooked because either it is not considered or because of technical inexperience in the identification of amoebas in stool or anoscopic aspirates.‘-” A rectal biopsy taken for diagnostic purposes may in this situation provide the only clue to diagnosis.“-” No systematic comparison of different staining methods for identification of amoebic trophozoites in tissue, however, has been made. In this study we compare five different staining methods on rectal biopsies from patients with amoebic dysentery, both befbre and after metronidazole therapy, and in paCents with bacillary dysentery.

Materials

and Methods

The study was performed in two stages. Initially, a randomized single-blind study of five different histologic stains previously used for detecting E. histolytica in tissue was carried out. Later, other conventional and immunologic techniques for the detection of E. histolytica were studied.

436

GILMAN

GASTROENTEROLOGYVol.78,No.3

ETAL.

Patients Patients were all adult males admitted to the Cholera Research Hospital (Dacca, Bangladesh) with clinical dysentery. Amoebiasis was diagnosed by finding hematophagous trophozoites of E. histolytica on microscopic examination of feces. Patients were diagnosed as having bacillary dysentery if: (a) stool microscopy revealed many fecal pus cells, and (b) shigella was isolated from a rectal swab. One patient with idiopathic ulcerative colitis was also included in the study. All patients on admission had proctoscopy and a rectal biopsy performed through a Welch-Allyn, T-cm anoscope; this was repeated at periods ranging from several hours to 3 wk after admission. If mucus was present, an attempt was made to include some with the biopsy specimen. No bowel preparation was given before anoscopic examination. Proctoscopic appearances were graded into mildlobular, or necrotic ulceration by criteria previously described.’ Rectal biopsies studied were obtained from amoebic and shigella dysentery patients both before and after appropriate therapy. Biopsies Rectal biopsies from patients were flattened (cut surface down) and pinned to small pieces of cork before fixing. Specimens were fixed in 95% ethanol, oriented on edge in paraffin, and processed by the Sainte-Marie technique.’ Sections (5-7 pm) were cut perpendicular to the mucosal surface, and 3-4 sections from different portions of the paraffin block were placed on the same slide. Between 15 and 20 slides from each block were made, and one slide was picked at random and stained by one of the following methods: (a) rabbit anti-E. histolytica serum conjugated with fluorescein isothiocyanate (direct technique); (b) Hematoxylin and Eosin (H & E); (c) a modified Periodic Acid Schiff (PAS)S; (d) Mallory’s Trichrome; and (e) Phosphotungstic Acid Hematoxylin (PTAH).” All sections were coded. Sections stained by the fluorescent antibody (FA) technique were read by one of us (R. Gilman); the other four stains were examined separately by two of us (R. Gilman and M. Islam) for the presence of amoebas. After completion of the initial reading, results of the different observers were compared, and a difference in the identification of amoebas occurred in 24 of 224 readings. These 24 slides were recoded and read again by the same two people. On reexamination there was disagreement only on one biopsy. The “corrected readings” were taken as the final readings. The patient identity code was then broken. The diagnoses of the patients were then correlated with the results of the slide readings. In the second part of the study, Gridley’s stain,” Whipf’s modified polychrome,” and a silver carbonate stain” were used on formalin-fixed tissue. An indirect immunoperoxidase technique,lz using rabbit antiamoebic sera, was performed on formalin-fixed tissue. Background staining was reduced by including the modifications suggested by Reading.13 Finally, an indirect FA test for E. histolytica was performed on ethanol- and formalin-fixed tissue. Formalin-fixed biopsies were obtained from a different patient group than were ethanol-fixed biopsies,

although some overlapping occurred. Both ethanol- and formalin-fixed specimens were coded before the indirect FA test was read.

Specific

Technical

Procedures

Direct fluorescent antibody test. One-half cubic centimeter of commercially obtained axenic amoebic (E. histolytica) antigen (ICN, Portland, Ore.) containing 1 x 10~ amoebas/ml mixed with complete Freund’s adjuvant was injected into two 1.5-kg New Zealand rabbits; two such injections, spaced 3 wk apart, were followed by a third injection without Freund’s adjuvant given intravenously 2 wk later. Rabbits were bled and sera were collected 3 wk after the last injection. This sera had a titer of 1:20,000 against E. histolytica antigen as determined by the indirect tanned red cell hemagglutination technique.14 These antiamoebic sera were conjugated with fluorescein isothiocyanate by the dialysis method of Clark and Shepard.” The conjugated sera were absorbed with acetone-dried rat liver powder and had a fluorescein protein (F/P) ratio of 3.4. As a control in the direct FA test, conjugated amoebic antisera were absorbed with 0.2 cm3 of amoebic antigen and 0.1 cm3 of phosphate-buffered saline (PBS), and flltered through a 0.45~pm Millipore filter. Tissue sections were tested for the presence of amoebas by applying conjugated antiamoebic sera for 30 min at a l/4 dilution, washed in PBS, and then mounted in glycerol. In control slides, absorbed conjugated antisera were applied for 45 min, washed in PBS, and mounted in glycerol. Amoebas did not stain when absorbed sera were applied. Indirect fluorescent antibody test. The indirect (FA) test for amoebas was performed using unconjugated rabbit antiamoebic sera. This sera were applied to tissue sections at a 1:4 dilution for 30 min, washed with three changes of PBS, and then the sections were incubated with fluorescein-conjugated goat antirabbit sera (Hyland, Costa Mesa, Calif.) at a 1:10 dilution for 30 min. After washing the slide again in PBS, Evans blue, a counter stain, was applied on ethanol-fixed tissues at a 1: 5,000 dilution: Because of a bright green, nonspecific tissue fluorescence, Evans blue was applied at a concentration of 1: 1200 on formalin-fixed tissue sections. Controls included slides in which normal rabbit serum or phosphate buffer were substituted for amoebic antisera. Also, as a control, nonconjugated amoebic antisera were absorbed with amoebic antigen (as described above) applied to tissue sections for 45 min, washed in PBS, and then traced with conjugated goat antirabbit sera. Unlike the direct FA test, absorbed sera in the indirect test decreased background staining, but did not abolish the staining of amoebas. All slides were coded before reading.

Results As seen in Table 1, the FA technique was the most sensitive technique (84%) to detect trophozoites in rectal biopsies from patients with amoebic dysen-

March 1980

Table

1.

437

IN RECTAL BIOPSY

E. histolytica

Identification of Amoebas in Ethanol-Fixed Rectal Biopsies Stained by Five Different

Techniques”

Number of biopsies positive for amoebas No. studied Acute amoebic dysentery before treatment Nonamoebic colitis patients P-value (Chi-square) ” FA = Fluorescent

antibody

19 a

Direct FA No. +/total

PAS No. +/total

H&E No. +/total

16 (84%) 0

14 (74%)

11 (58%)

0 >a.1

test; PAS = Periodic acid Schiff: Trichrome

Trichrome No. +/total

PTAH No. +/total

11(58%)

0 < 0.1

10 (53%)

0 < 0.1

= Masson’s trichrome

0 < 0.05

stain; PTAH = Phosphotungstic

acid

hematoxylin.

tery (P-value of < 0.05 compared to PTAH, and P < 0.10 compared to Trichrome and H & E stains). Of the permanent stains, the PAS technique was most sensitive. Amoebas could be identified rapidly and easily at low power (10X) by the fluorescent antibody technique and PAS stain. The other three techniques (H & E, PTAH, and Trichrome) required considerably more time to find amoebas in tissue sections. In 15 patients (16 tissue blocks) with amoebiasis treated for over 24 hr with metronidazole, or in 8 patients with shigellosis, there were no falsepositive readings with either the direct FA technique or any of the four permanent staining techniques. Nuclear detail was very helpful in specific identification of E. histolytica, but this was obscured in amoebas stained with PAS. Nuclear detail was made more distinct by overstaining with freshly prepared Weigert-Lillies hematoxylin. However, PAS-stained amoebas had a characteristic vacuolation of their cytoplasm, making accurate identification possible. Amoebas invading tissue were usually found in aggregations that were present only in areas of necrosis and tissue disruption. All 10 patients with necrotic ulcers seen on proctoscopy had amoebas detected in their rectal biopsies, but in 3 of the 9 patients with a mild-lobular proctoscopic appearance, the rectal biopsy was negative for amoebas by all techniques. Other staining procedures-Gridley’s stain, silver carbonate, and Whipf’s modified polychrome stain -did not stain amoebas more distinctly than did either PAS or Trichrome stain. Ten amoebas-positive and 5 amoebas-negative ethanol-fixed sections examined by the indirect fluorescent antibody technique gave the same results as obtained by the direct FA technique. Although amoebas stained more brightly by the indirect method, this advantage was minimized by increased background staining. Formalin-fixed tissue studied by the indirect fluorescent antibody technique was quite specific. As seen in Table 2, only one false-positive result was recorded. The major problem was high background fluorescence. This was decreased only when lower-

than-usual dilutions of Evans blue were used. Amoebas in formalin-fixed tissue were identified more frequently (not statistically significant) by the indirect FA test than after staining with H & E or PAS stains (Table 2). There was a discrepancy of 74% vs. 44% detection of E. histoiytica by the PAS technique in ethanol compared with formalin-fixed rectal biopsies (Tables 1 and 2). Immunoperoxidase staining of formalin-fixed tissue sections was not as reproducible as the fluorescent antibody technique. Background staining of necrotic tissue and red cells (areas where amoebas are present) was difficult to eradicate and interfered with trophozoite identification. Three of 6 patients treated with metronidazole 16 hr before rectal biopsy, had faintly staining E. histolytica trophozoites visable by direct FA technique, but not by conventional stains. Amoebas were not seen by any staining technique in biopsies taken from patients who had received metronidazole for 24 hr or longer.

Table 2.

Identification of Amoebas in Formalin-Fixed Rectal Biopsies Detected by the Indirect Flourescent Antibody Technique Compared Results with PAS and H & E Stains

Number E. histolytica patients before metronidazole therapy E. histolytica patients 24 hr or more after metronidazole therapy Shigella patients a No significant

18

17 5

Indirect fluorescent” antibody

13 (72%)b

0 1 (ZO%)C

PAS

8 (44%)

0 0

to

H&E

10 (56%)

0 0

differences were found between the three staining techniques. b Four of the five missed biopsies were from patients with mild or lobular proctoscopic appearance. One patient had necrotic ulcers. Of the 13 positive patients, 11 had necrotic ulcers and 2 had mild-to-lobular proctoscopic appearance. c This was a biopsy from a patient with severe shigellosis and is a false-positive result.

438

GILMAN

ET AL.

Discussion In this study, the fluorescent antibody technique was compared with other methods for identification of amoebas in rectal biopsies. Although equally specific, the FA technique was more sensitive in detecting E. histolytica trophozoites than routine histologic stains. Two previous studies have identified E. histolytica trophozoites in tissue using the indirect FA technique.16.” Both studies used formalin-fixed tissue to which human antiamoebic sera were applied. The antiamoebic sera were then traced by an antihuman globulin-fluorescein conjugated antisera. We found marked exudation of human IgG into the lamina propria and onto the surface of the coionic epithelium in intestinal amoebiasis (personal observation). Edema, exudation, and focal hemorrhage are often found in rectal biopsies from patients with amoebic dysentery.‘” High levels of immunoglobulin (Ig) present in human biopsies associated with inflammatory changes may produce marked nonspecific background fluorescence when antihuman globulin-fluorescein conjugated sera are used for tracing. This fluorescence will be further intensified by the autofluorescence present in formalin-fixed tissues. To decrease background fluorescence we modified the previously described FA technique in two ways. First, we followed the Sainte-Marie technique, which uses ethanol fixation; this decreases background staining” without impairing the staining of amoebas by conventional stains. Secondly, we used antisera made in rabbits, which can be traced with a FA technique that does not react with human Ig. Rabbits challenged with a commercially available axenic amoebic antigen provide a convenient and reproducible source of antiamoebic sera. Rabbit antiamoebic sera were used both for the direct (after conjugation) and the indirect (uncon jugated) fluorescent labeling. Both methods were sensitive and specific in ethanol-fixed tissues; however, background fluorescence was somewhat greater with the indirect FA technique. The indirect FA technique may be more convenient, since antirabbit fluorescein-conjugated antisera are available commercially. The indirect FA technique was also tried on formalin-fixed tissue with moderate success. Background staining was higher and amoebic staining less distinct than in ethanol-fixed tissue. Even with these disadvantages the indirect FA method used on formalin-fixed tissue was reasonably specific and can be performed with commercially conjugated antisera. It permits rapid scanning and identification of amoebas on routinely processed biopsies.

GASTROENTEROLOGY

Vol. 78, No. 3

The PAS technique was only slightly less sensitive than the FA technique in detecting amoebas, but was specific. Amoebas are stained by PAS because of their glycogen content. The discrepancy in E. histolytica detection of 74% vs. 44% ifi ethanol compared to formalin-fixed, PAS-stained biopsies, may be due to the increased solubility of glycogen in a water-based fixative. The permanence of the stained preparation is an advantage ,in the use of this technique. The modified PAS technique we used has the advantages of simplicity in its preparation, stability of its reagents even in a tropical environment, and decreased background staining. In both the direct and indirect FA technique as well as in the PAS technique, amoebas could be detected by scanning at lower power. This was not true of the other conventional techniques studied. Difficulty identifying amoebas in exudate in H & E sections when low magnification is used has been previously described.‘” The immunoperoxidase method was not an acceptable substitute for detecting amoebas in tissue because of the large amount of nonspecific background staining present in areas of necrosis. In this study the percentage of patients correctly diagnosed by rectal biopsy was higher than that found in several previous studies.7.‘9.20First, because (as suggested by Juniper et al.“), an attempt was made to include mucus with the biopsy. Secondly, 10 of 19 patients studied had necrotic ulcers seen on proctoscopy. That identification of amoebas in rectal biopsies from patients with necrotic ulcers is higher than in patients having a mild-lobular rectal appearance,7 was confirmed in this study. All patients with amoebic dysentery, whose diagnosis was missed by the direct FA techniques, had a mild-to-lobular proc toscopic grading. It should be emphasized that the failure to demonstrate E. histolytica in mucosal biopsies from dysentery patients is not evidence that the diagnosis of amoebiasis has been excluded.‘g Using both FA and conventional staining techniques, we were unable to confirm reports that amoebas persist in tissue in altered morphologic forms after therapy.“’ Rectal biopsies are not necessary for the routine diagnosis of amoebic dysentery. Examination of feces and proctoscopic aspirates by an experienced observer is a rapid, easy, and effective means of diagnosis. Rectal biopsy, in fact, is a relatively inefficient method for the diagnosis of amoebic colitis. However, amoebiasis, especially in developed countries with a temperate climate, may be missed due to the inexperience of laboratory workers in the identification of E. histolytica or because the diagnosis of amoebiasis has not been considered. We would suggest that in all cases of colitis, sec-

March

1980

E. histolytico

tions from rectal biopsy specimens be routinely stained with PAS. Optimally, where facilities for fluorescent microscopy are available, it is advantageous to fix rectal biopsies in ethanol. Paraffin sections may then be examined by either the direct or indirect FA technique for amoebas as well as by the more routine H & E and PAS stains. In the absence of ethanol fixation, formalin-fixed tissue may be substituted.

9. Horobin

10. 11.

12.

13.

References 14. 1. Krogstad

2.

3.

4. 5. 6.

7.

8

DJ, Spencer HC, Healy GR, et al: Amebiasis: Epidemiologic studies in the U.S. 1971-1974. Ann Int Med 88:8997, 1978 Wruble LD, Duckworth ]K, Duke DD, et al: Toxic dilatation of the colon in a case of amebiasis. N Engl J Med 275:926-928, 1966 Tucker PC, Webster PD, Zachary M, et al: Amebic colitis mistaken for inflammatory bowel disease. Arch Intern Med 135:681-685, 1975 Krogstad DH, Spencer HC, Healy GR: Current concepts in parasitology-amebiasis. N Engl J Med 298262-265, 1978 Juniper K jr, Steel VW, Chester CL: Rectal biopsy in the diagnosis of amebic colitis. South Med J 51:581-587, 1958 Goodman MJ, Kirsner JB, Riddel RH: Usefulness of rectal biopsy in inflammatory bowel disease. Gastroenterology 72:952-956, 1977 Gilman RH, Prathap K: Acute intestinal amoebiasis-proctoscopic appearances with histopathological correlation. Ann Trop Med Parasitol 63:359-365, 1971 Sainte-Marie G: A paraffin embedding technique for studies employing immunofluorescence. J Histochem Cytochem 10:250-256,

1962

15. 16.

17.

18.

19.

20. 21.

IN RECTAL

BIOPSY

439

RW, Kevill-Davies IM: Basic fuchsin in acid alcohol: a simplified alternative to Schiff reagent. Stain Technol46:5358, 1971 Preece A: A Manual for Histologic Technicians. Third edition. Boston, Little Brown and Co., 1972 Wilson R, Fairchild DG: Modified Whipf’s polychrome: a connective tissue stain with special application for demonstrating leishmania. Stain Technol 52:105-111, 1977 Culbcrtson C: Soil ameba infection-specific indirect immunoenzymatic (peroxidase) staining of formalin-fixed paraffin sections. Am J Clin Path01 63:475-482, 1975 Reading M: A disgestion technique for the reduction of background staining in the immunoperoxidase method. J Clin Path01 30:88-90, 1977 Kessel JF, Lewis WP, Molina P, et al: Indirect hemagglutination test and complement fixation tests in amebiasis. Am J Trop Med Hyg 14:540-550, 1965 Clark HF, Shepard CC: A dialysis technique of preparing fluorescent antibody. Virology 20:642-644, 1963 Parelkar SN, Stamm WP, Hill KR: Indirect immunofluorescent staining of Entamoeba histolytica in tissues. Lancet 1:212-213, 1971 Hoffman EO, Miller MJ: Immunofluorescent staining of amebae in routine paraffin-embedded tissues. J Parasitol 61:1104-1105, 1975 Pittman FE, El-Hashimi WK, Pittman jC: Studies of human amebiasis. II. Light and electron-microscopic observations of colonic mucosa and exudate in acute amcbic colitis. Gastroenterology 65:588-603, 1973 Pittman FE, Hennigar GR: Sigmoidoscopic and colonic mucosal biopsy findings in amebic colitis. Arch Path01 97:155158, 1974 Prathap K, Gilman R: The histopathology of acute intestinal amebiasis. Am J Path01 60:229-245, 1970 Doxiades TH, Yiotsas HZ: Observations, findings and conclusions regarding proctosigmoidoscopy and colon biopsy. Am 1 Proctol 16:361-369, 1965