Comparison of Assays for Antibodies To

Comparison of assays for antibodies toEncephalitozoon cuniculi in rabbits

R. Boot1, A. K. Hansen2,3, C. K. Hansen4, N. Nozari' & H. C. W. Thuis1

'Section of Laboratory Animal Microbiology, LIS, National Institute of Public Health and theEnvironment (RIVM), Bilthoven, The Netherlands, 2Department of Experimental Medicine, The PanumInstitute, University of Copenhagen and National University Hospital, Denmark, 3Division ofLaboratory Animal Science and Welfare, Department of Pharmacology and Pathobiology, RoyalVeterinary and Agricultural University, Frederiksberg, Denmark and 4Mathematics Department,Eastern Washington University, Cheney, Washington, USA

SummaryTwo indirect immunofluorescence (IIF)assays, two enzyme-linked immunosorbent assays(ELISAs)and the carbon immunoassay (CIA)for determination of antibodies toEncephalitozoon cuniculi were compared using 210 sera of rabbits, 135 of which originatedfrom seven infected colonies, while 75 originated from four uninfected colonies. There wasnoevidence of a difference between the different assays with respect to the number of positivesera. There was a clear correlation between the quantitative response measured by IIF andCIA and the other assays, and between both IIF tests, while no such correlation was found inthe quantitative response measured by ELISAs,which might be explained by the lessquantitative nature of the ELISA.Therefore quantitative determination of antibodies to E.cuniculi should be performed by IIF and not by ELISA.The nosographic sensitivities Nt andspecificities N2 of the assays were ~ 0.94 and ~ 0.97 respectively. Small differences in Ntand N2 between the assays, although not statistically significant, were responsible fordifferences in the calculated predictive values of a positive test and of a negative test. Asexpected, the magnitude of these differences depended on the fraction of positive sera sampledfrom a given colony. There was strong evidence of such a difference between the fraction ofpositive sera found in different colonies, but the sample size from some colonies was toosmall to allow any conclusion, whether this was due to differences in the prevalences of theinfection in the colonies or something else. We conclude that any of the assays will besuitable for the routine health monitoring of laboratory rabbit colonies for E. cuniculiinfection, as recommended by the Federation of European Laboratory Animal ScienceAssociations.

Keywords Encephalitozoon cunciculi; serology; health monitoring; rabbits

The microsporidian intracellular parasite E.cuniculi infects a wide range of hostsincluding the common small laboratory ani-mal species and human and non-human pri-mates (Canning & Lorn 1986, Canning &Hollister 1992).E, cuniculi has recentlyemerged as an opportunistic parasite in

Correspondence to: R. Boot

patients infected with the human immuno-deficiency virus (Desplazes et al. 1996).Whether E. cuniculi is a homogeneous spe-cies is currently unclear: phenotypic andgenotypic differences between isolates fromdifferent hosts have been found, but generallyisolates are infective to hosts other than theoriginal host species upon experimentalinoculation (Didier et al. 1995).

Accepted 3 March 2000 © Laboratory Animals Ltd. Laboratory Animals (2000) 34, 281-289

at UNIV AUTONOMA DE NUEVO LEON on January 23, 2015lan.sagepub.comDownloaded from

http://lan.sagepub.com/

282

In rabbits E. cuniculi infection usually cau-ses a mild, subclinical disease, presumablyaffecting the animals' immune responsiveness(Cox 1977),but the infection can also manifestitself as an overt and sometimes fatal disease(Canning &. Hollister 1992, Pakes &. Gerrity1994).The Federation of European LaboratoryAnimal Science Associations (FELASA)recommends the periodic monitoring oflaboratory colonies of rabbits for E. cuniculiinfection (Kraft et a1. 1994).

Encephalitozoon cuniculi infections havetraditionally been diagnosed by histologicalexamination of granulomatous lesions inbrain and kidneys (Pakes &. Gerrity 1994).The application of molecular techniques(Fedorko &. Hijazi 1996) to detect the parasitein laboratory animals has not been described.Immunologically based screening proceduresthat have been applied in rabbits consist of anintradermal delayed type hypersensitivitytest (Pakes et a1. 1972) and various assays todetect antibodies to the parasite. The validityof antibody assays for the detection of E.cuniculi infection has been evaluated bycomparison with histology, and in rabbitsboth compared favourably (Waller 1977,Pakes et a1. 1984, Greenstein et a1. 1991).Some studies have compared serologicalassays, e.g. CIA and IIF (Greenstein et a1.1991, Waller 1997). The IIF has been com-pared with complement fixation (Pakes et a1.1984) and a microbead agglutination test(Shadduck &. Geroulo 1979, Pakes et a1.1984).Although the various serologicalassays were said to be equally useful, it islikely that the sample sizes were too limitedto draw firm conclusions, and/or statisticalanalysis of the data were not appropriate.

The use of ELISAhas not been reported inthe monitoring of rabbits for E. cuniculiinfection. As the latter assay was set up inboth of our laboratories (RIVMand UoC), wecompared our ELISAswith our IIFs and theCIA (UoC only) using rabbit sera sampled incolonies with and without previous histolo-gical evidence of E. cuniculi infection, inorder to evaluate whether any differencesbetween the assays with respect to theirsuitability for diagnosing this agent on acolony level, e.g. as recommended by theFELASA(Kraft et a1. 1994Lcould be

Laboratory Animals (2000) 34

Boot et al.

observed. Here the null hypothesis, definedas 'the assays are equally applicable forhealth monitoring', is tested against thealternative that 'the assays are not equallyapplicable for health monitoring'.

Materials and methods

SeraOur study was based on a sample of 135rabbits from seven colonies in which nose-miasis had been diagnosed by histology(Table 1; infected coloniesl, and a sample of75 rabbits from four colonies without such ahistory (uninfected colonies). In all coloniesbut one (K),some kind of barrier protectionand health monitoring was applied. In colo-nies E and J breeding rabbits were periodi-cally monitored for E. cuniculi infectionusing the ELISAperformed by UoC (seebelow), and seropositive animals were movedto colony 1.Colony K was a conventionalrabbit colony which was kept without spe-cific measures to protect the animals frominfection by pathogenic microorganisms.Sera were transported in a frozen conditionand stored at - 20°C until testing.

Serological assaysTesting of split samples was done by theSection of Laboratory Animal Microbiology,LIS,National Institute of Public Health andthe Environment, Bilthoven, The Nether-lands (RIVMLand by the Department ofExperimental Medicine, The Panum Insti-tute, University of Copenhagen and NationalUniversity Hospital, Denmark (UoC).

Indirect immunofluorescence test

RIVM The IIF was performed as described(Bootet a1. 1988).E. cuniculi was propagatedin RK 13 cell cultures, and spores werecounted using a haemocytometer underphase-contrast and adjusted to a concentra-tion of 1.4 x 109 spores/ml. Volumes of 2.5 IIIof the spore suspension were dropped perslide, air dried, fixed in methanol and storedat - 20°C before use. Slides were incubatedwith test sera for 45 min at 37°C, washed 3 xin PBSpH 7.2 at room temperature, stainedfor 30 min at 37°C with a FITC-conjugated



Encephalitozoon cuniculi antibody assays

Table 1 Comparison of assays for detection of antibodies to Encephalitozoon cuniculi in sera ofrabbits from colonies with and without histological evidence of nosemiasis

No. of positive samples

RIVM* Uoct

Colony status§ Colony No. of samples !IF ELISA IIF ELISA CIA % positive

Un infected A 10 0 0 0 0 0B 38 0 1 0 0 0C 7 0 1 0 0 0D 20 0 0 0 0 0

Subtotal 75 0 2 0 0 0Infected E 10 0 0 0 0 0 ...-0

F 4 4 4 4 4 4 ...-100G 10 8 9 8 9 7 70-90H 12 7 7 7 7 7 ...-58I 16 3 3 5 3 3 19-31J 55 2 3 2 2 2 4-5K 28 25 26 26 24 24 86-93

Subtotal 135 49 52 52 49 47All 210 49 54 52 49 47

*National Institute of Public Health, Nl (RIVM) and tUniversity of Copenhagen, DK (UoQ§It was attempted to free colonies E and J from the infection by moving all ELISApositive rabbits to colony IIIF= indirect immunofluorescence; ELISA= enzyme-linked immunosorbent assay;CIA= carbon immunoassay

283

goat anti-rabbit Ig (Dako Ltd, Glostrup,Denmark), washed again 3 x in PBS,air driedand examined using a fluorescence micro-scope at 200 x magnification. Sera werescreened at 1:10 dilution in PBSpH 7.2.Positive sera were re-examined using dou-bling dilutions up to 1:2560 in PBS.uoe The IIF was performed as described(Hansen et al. 1992)using antigen (Testman,Uppsala Swedenl, coated onto slides, airdried, flame fixated and stored at - 20°Cuntil use. Slides were incubated with testsera for 30 min at room temperature, washed3 x in PBSpH 7.2, stained for 30 min at roomtemperature with FITC-conjugated goat anti-rabbit Ig (Dako, Glostrup, Denmark), washedagain 3 x in PBS,air dried and examinedusing a fluorescence microscope at 400 xmagnification. Sera were screened at 1:20dilution in PBSpH 7.2. Positive sera were re-examined using doubling dilutions up to1:2560 in PBS.Enzyme-linked immunosorbent assay

RIVM The E. cuniculi used as the antigenoriginated from a spontaneously infected

rabbit. Stock suspension of E. cuniculi sporeswas diluted in Na2C03 buffer pH 9.6, anddilutions containing 1.4x 108 spores/ml werecoated for 24 h at 22°C to high bindingpolystyrene flat-bottomed micro titre plates(Greiner, Alphen aan de Rijn, The Nether-lands) and plates were stored at - 20°Cuntiluse. The ELISAwas performed as describedfor Streptobacillus moniliformis (Boot et al.1993):sera were diluted 1:50 in PBSpH 7.2containing 1% BSA,and 100 /-llvolumes wereincubated in duplicate with the antigen for1 h at 3rc and then treated with 100 !-II ofperoxidase-labelled goat anti-rabbit Ig con-jugate (Cappel, Malvern, USAl. Tetramethyl-benzidine in DMSO was used as substrateand the reaction was stopped after 10 min by2 mol/l H2S04. Absorbances were read usinga Titertek multiscan at 450 nm. Antigen andconjugate concentrations were pre-tested bychessboard titrations with various dilutionsof the antigen and the conjugate. Uncoatedcontrol wells exposed only to substrate andstopping reagent were used to blank thereader. The ELISAwas considered positive ifthe optical density (OD)value (extinction) of




284

the 1:50 serum dilution exceeded themean + 4 SD of the mean of the ODs in thesera of rabbits from colonies without a his-tory of E. cuniculi infection. The aD valuesof 2 samples that were evidently falselypositive by the RIVM ELISA were not inclu-ded in the calculation of cut-off values.UoC The ELISA was performed as pre-viously described (Hansen & Skovgaard-Jen-sen 1995): antigen-coated microtitre plates(Charles River Wiga, Sulzfeld, Germany)were stored at - 20°C until use. Sera werediluted 1:100 in 0.1 % PBS-tween 20 and200 Jll volumes were incubated with theantigen and with uncoated wells for 2 h atroom temperature, and after washing weretreated with 200 III of peroxidase-labelledgoat anti-rabbit Ig conjugate (Dako Ltd,Glostrup, Denmark) for 2 h at room tem-perature. O-phenyldiamin was used as sub-strate and the reaction was stopped after10 min by 2 mol/l H2S04. Absorbances wereread at 492 nm using a Sigma microplatereader (Sigma, St Louis, USA). Antigen andconjugate concentrations were pre-tested bychessboard titrations with various dilutionsof the antigen and the conjugate. Uncoatedcontrol wells exposed only to substrate andstopping reagent, were used to blank thereader. The ELISA was considered positive ifboth the difference between the ODs of thecoated and uncoated well was more than 0.2,and if the OD of the coated well was morethan 0.394 (the mean + 3 SD of a range of seraof rabbits from colonies without a history ofE. cuniculi infection).

Carbon immunoassayThe assay was performed as described (Waller1977) using antigen (Testman, Uppsala,Sweden), coated onto slides, air dried, flamefixated and stored at - 20°C until use. Serawere screened at 1:20 dilution in PBS pH 7.2.Five microlitre of the diluted serum wasmixed with 5 III india ink (Testman, Uppsala,Sweden) and left on the coated slide for5 min. Slides were then washed with PBS toremove all ink and were finally examined foragglutination of spores using a light micro-scope at 400 x magnification. Positive sera


Boot et al.

were re-examined using doubling dilutions ofup to 1:2560 in PBS.

In all assays (ELISA, IIF and CIA) negativeand positive control sera (the latter obtainedfrom rabbits immunized with spores of E.cuniculi) were run in each test.

Evaluation of data/statistical analysisDifferences in the numbers of positive sam-ples obtained using the assays and in thenumbers of seropositive rabbits betweencolonies were analysed using the Chi-squaretest. Multiple correlation coefficients of thecorrelation of the quantitative response ofeach assay as compared with the other tests,as well as correlation coefficients of the cor-relation of the quantitative response of RIVMELISA to UoC ELISA, and RIVM IIF to UoCIIF were calculated. The sensitivity and spe-cificity of the assays were calculated as thenosographic sensitivity (NIl and the noso-graphic specificity (N2), respectively. N1 isdefined as the proportion of infected rabbitsreacting truly positive (TP) in the test and N2

as the proportion of uninfected rabbitsreacting truly negative (TN) in the test(Hansen 1993). In order to estimate N1 andN2, serum samples were considered TP or TNupon showing three positive or negativeresults, respectively, after testing by the fiveassays. Differences in the sensitivity andspecificity of the assays were evaluated bythe Chi-square test. The predictive values fora positive test (PV + I and a negative test(PV - )were calculated according to Jacobsonand Romatowski (1996):

PV+ = P X N1P X N1 + (1 - P) x (1 - N2)

PV- = (1- P)X N2

(1 - P) x N2 + P x (1 - N1)

in which P =prevalence of infection (= % ofpositive sera), N1 =nosographic sensitivityand N2 =nosographic specificity of the assay.

ResultsTable 1 shows the outcome of determina-tions of antibodies in rabbit sera to E. cuni-culi using IIF, ELISA and CIA. Sera of rabbitsfrom the four colonies not having a history of




Table 2 Chi-square tests for comparing differences in number of positive samples betweencolonies and between assays

Positive samples in the colony

285

Colony

Total

Positive observed

ExpectedNegative observed

Expected

l = 76.5; df = 5; P = 0.000

F

441.57

o2.43

G

1083.92

26.08

H K All

12 16 55 28 125

7 3 2 25 49

4.70 6.27 21.56 10.98

5 13 53 3 767.30 9.73 33.44 17.02

Positive samples found in different tests

RIVM

IIF ELISA

Total 125 125Positive observed 49 52Expected 49.8 49.8Negative observed 76 73Expected 75.2 75.2

/=0.63; df=4; P=0.96. See Table 1 for abbreviations

UoC

IIF ELISA CIA All

125 125 125 625

52 49 47 249

49.8 49.8 49.8

73 76 78 376

75.2 75.2 75.2

See Table 1 for abbreviations

Table 3 Differences in test results in 12 samples ofrabbits from Encephalitozoon cuniculi infected colo-nies using various antibody assays

invariably increased to 100% but the PV+values decreased for most assays, with theexception of the RIVM IIF (Table 4).Thenosographic sensitivity (NIl of the assays(Table 4) was at least 0.94 and the IIFper-formed by UoC showed maximal sensitivity

E. cuniculi infection were negative in allassays, except two (2.7%) that were clearlypositive in the RIVM ELISA(Table 1).Allassays detected positive sera in six of theseven colonies with previous histologicalevidence of E. cuniculi infection. Applyingthe X2 test (Table 2) there is strong evidenceof a difference between the fraction of posi-tive sera found in the different colonies,while there is no evidence of a difference inthe number of positive samples detected bythe different assays.

In 12 of the 135 (8.9%) sera from infectedcolonies the outcome of the assays disagreed(Table 3), but no systematic differences werefound and all assays, except the RIVMIIF (seebelow), yielded one or more apparently aber-rant results.

Table 4 shows the predictive values of apositive test (PV+)and a negative test(PV- ) calculated from Nl and N2 for theassays, in colonies showing (arbitrarily) high,intermediate and low prevalences. The PV+and PV - were all above 93% in coloniesshowing an intermediate (30%) or high (60%)prevalence of infection. With low pre-valences (10 to 1%) the PV - almost

Colony

G

K

RIVM UoC

Sample IIF ELISA IIF ELISA CIA

1 +2 + +3 + + + +4 +5 +6 +7 +8 + + + +9 +10 + +11 + + + +12 + + +




286

Table 4 The nosographic sensitivity (N,)* and specificity (N2) of five antibodyassays for detection of Encephalitozoon cuniculi infection in rabbits andcalculated predictive values of positive (PV +) and a negative test (PV - ) atsome assumed prevalences

RIVM UoC

!IF ELISA IIF ELISA & CIA

Sensitivity (N,)* 0.98 0.98 1.0 0.94Specificity (N2) 1.0 0.97 0.98 0.99

Prevalence %

PV+ 60 100 98.0 98.7 99.330 100 93.3 95.5 97.610 100 78.4 84.7 91.21 100 24.8 33.6 48.7

PV- 60 97.1 97.0 100 91.730 99.2 99.1 100 97.510 99.8 99.8 100 99.31 100 99.9 100 99.9

*N, and N, are the proportions of infected rabbits reacting positive or negative respectivelyin ;;. 3 assays.See Table 1 for abbreviations

Boot et al.

(1.0). The nosographic specificity (N2) was atleast 0.97 and the IIF performed by RIVMshowed maximal specificity (1.0).There wereno statistically significant differences in thesensitivity or specificity of the assays Iitest,pNl > 0.5, pN2 > 0.25).

Table 5 shows the multiple correlationcoefficients of the quantitative responsemeasured by each test as compared with theother tests, while Table 6 shows the correla-tion coefficients of the quantitative outcomeof RIVM ELISAto UoC ELISA,as well asRIVM IIF to UoC IIP.No correlation wasfound in the antibody levels measured byELISAs,neither when correlated to the othertests nor to one another, while there wasclear correlation between the quantitative

outcome of IIF and CIA and the other tests,and between the two IIF tests.

DiscussionThe overall results of serological monitoringof the rabbit colonies (Table 1)agreed verywell with their history. Those four colonieswithout previous histological evidence of E.cuniculi infection (considered as uninfected)were seronegative in all but one assay. It isunclear why two sera from two uninfectedcolonies were clearly positive in the RlVMELISA(Table 1).All but one of the seveninfected colonies showed seropositive ani-mals in all assays. Colony E appeared nega-tive upon testing, suggesting that systematic

Table 5 Multiple correlation coefficients of the correlation of antibody levels toEncephalitozoon cuniculi measured by each of five different assays as comparedwith the other assays (49 samples)

RIVM UoC

ELISA IIF ELISA !IF CIA

Multiple c.c. adjusted for 1.5 55.4 0.0 50.7 54.8degrees of freedom (r) %

P value* 0.331 0.000 0.625 0.000 0.000Power 0.0512 0.9671 0.0500 0.9354 0.9639

*Analysis of variance, using ho: D =O. SeeTable 1 for abbreviations




Encephalitozoon cuniculi antibody assays 287

*Analysis of variance, ho: D=O. See Table 1 for abbreviations

Table 6 Correlation of the antibody levels measuredby RIVM ELISA to UoC ELISA. as well as RIVM IIF to UoCIIF (49 samples)

removing of all ELISA-positivebreeders hadfreed the colony from infection. Although theperiodic testing of all animals and theremoving of all positive responders has beenclaimed to be successful in the eradication ofE. cuniculi infection from rabbit colonies(Cox et al. 1977,Bywater &. Kellett 1978,Waller 1998),the number of samples in thiscolony, however, was too small to draw firmconclusions about its infection status.

It must be kept in mind that samplingeight animals from a colony (as is recom-mended by the FELASA(Kraft et al. 19941)will only allow detection of infections if theyare present in about 30% of the animals (with95% probability) and prevalences of less than30% seem to be common in rabbits (Waller1988).In this study, there is strong evidenceof a difference between the number of posi-tive sera found in the different colonies(Table2).Whether this is due to the fact thatthere really is a difference in prevalence of .the infection or something else cannot bestated, due to the small sample size of somecolonies. There is no evidence of a differencein the number ofpositive sera detected by thedifferent assays (Table2). A sample size of atleast 3100 would have been necessary todetect a difference between proportions asclose as those observedwhen the power is setto 0.90. However, such small differences areof no importance for the application of theseassays in routine health monitoring.

It is not clear why 12 samples from infec-ted colonies showed different outcomes inthe assays (Table3). Differences in the out-come of IIFsand ELISAsvs CIA may be dueto the fact that IIF and ELISAmeasure IgMand IgG (due to the specificity of the con-jugates usedl, whereas the CIA measures 7SIgG antibody only (Waller et al. 1978).If the

Correlation coefficient adjustedfor degrees of freedom (r) %

Pvalue*Power

ELISA IIF

0.6 43.6

0.259 0.0000.0502 0.8486

E. cuniculi infection exists for some time,IgG antibodies will develop together withhistological lesions, and it is therefore notsurprising that histologically positive rabbitswere invariably seropositive by IIF and CIA(Waller 1977, Greenstein et al. 1991). Inher-ent to our study the time of infection of therabbits is unknown. The data given by Pakeset al. (1984)cannot be compared with ours astheir sample size was very limited and part ofthe samples were pooled sera. Recent infec-tion with IgM antibodies only, may havebeen present in the CIA-negative rabbits thatwere positive in at least one of the otherassays. Ig class differences, however, cannotexplain a CIA-positive outcome in a samplethat was negative in all other assays nor canthey explain variable results in the IIFs andELISAs(Table 3). Differences between theassays may have occurred due to differencesin the antigen. However, studying these 12samples with a non-uniform outcome, theUoC-IIF and the UoC-CIA, which are basedupon the same antigen, are inconsistent withthe other tests, and therefore there is noreason to believe that the differences are dueto differences in the antigen.

If the quantitative response measured bythe ELISAsis correlated to the other tests orto one another, the correlation is too small tobe of any interest (Tables 5 and 6). To be ofany interest the correlation should be morethan 50%, which could have been shown inthe sample size of 49 with a power of morethan 0.90, but no such correlation was found.There is, however, correlation between thequantitative response measured by IIF andCIA and the other tests and between the twoIIF tests (Tables 5 and 6). This may beexplained by the fact that IIF and CIA titresare determined in serially diluted sera, i.e.there is no upper limit in contrast to ELISAwhere an upper limit is determined by thequantity of antigen coated to the well. Whenthere is no more antigen available for anti-body to react with, the washing step willremove excess antibody. Therefore, quanti-tiative examinations should be performed byIIF and not by ELISA.

No serological assay can identify withabsolute certainty which rabbit is infected. Inorder to compare the sensitivity and the spe-




288

cificity of the assays we considered a serum tobe true positive (TP) or true negative (TN)upon showing three positive or negativeresults respectively using the five assays.Subsequently we calculated the nosographicsensitivity (Nd, defined as the proportion ofinfected rabbits reacting as TP in the test, andthe nosographic specificity (N2J as the pro-portion of uninfected rabbits reacting as TNin the test (Hansen 1993).This approach didnot reveal that any differences in the sensi-tivity and specificity of the assays were of anyinterest in practice (Table 4).

As expected, the nosographic sensitivity(Nd and specificity (N2) of the assaysdetermined the predictive values of a positivetest (PV+ J as well as that of a negative test(PV - ) (Table 4). As the prevalence of infec-tion (or the % of seropositivesl drops, thePV - increases to very high levels, even forassays with moderate N1 and N2 values (Tyler& Cullor 1994J.Simultaneously the PV+falls precipitously for assays in which thespecificity (N2) is not maximal (1.0); this wasthe case with all assays with the exception ofthe RIVM 1IF. Given their high sensitivity, allassays seem to be suitable for screening rab-bit colonies for E. cuniculi infection. Positivetest results obtained with assays other thanthe RIVM IIF might be confirmed by thelatter assay, as its specificity appeared max-imal, so all its positive results might betrusted. It should be mentioned that themethod applied for the determination of trueand false positives implies that some of thecalculated specificities might be too low asall false positives-except those two found bythe RIVM ELISA-were found in colonies inwhich infection was actually present.

A larger sample might have resulted in astatistically significant difference betweenthe specificities of e.g the RIVM IIF and theRIVM ELISAor the DoC IIF and the DoCELISA/CIA. However, a specificity of 0.97 isacceptable for an assay applied in laboratoryanimal health monitoring, and a lower sen-sitivity can be fully compensated by sam-pling more animals in a random sample forroutine monitoring (Hansen 1993J.Wetherefore conclude that IIF,ELISA and CIAare all suitable for the routine health mon-itoring of laboratory rabbit colonies for E.


Boot et al.

cuniculi infection as recommended byFELASA(Kraftet a1. 1994).The occurrence ofsmall differences in sensitivity and specifi-city of the assays performed in differentlaboratories must however be kept in mind,as they may have a considerable impact onthe predictive value of both positive andnegative test results in colonies with a lowprevalence of infection.

References

Boot R, van Knapen F, Kruijt BC, Walvoort HC (19881Serological evidence for Encephalitozoon cuniculiinfection (nosemiasis) in gnotobiotic guineapigs.Laboratory Animals 22, 337-42

Boot R, Bakker RHG, Thuis H, Veenema TL,de HoogH (1993) An enzyme-linked immunosorbent assay(ELISA) for monitoring rodent colonies for Strep-tobacillus moniliformis antibodies. LaboratoryAnimals 27, 350-7

Bywater TEC, Kellett BS (19781 The eradication ofEncephalitozoon cuniculi from a speCific pathogenfree rabbit colony. Laboratory Animal Science 28,402-4

Canning ED, Lom T (1986) The Microsporidia ofVertebrates: New York: Academic Press

Canning ED, Hollister WS (19921 Human infectionswith microsporidia. Review of Medical Microbiol-ogy3,34-42

COX TC (1977) Altered immune responsiveness asso-ciated with E. cuniculi infection in rabbits. Infec-tion and Immunity 15, 392-5

COX JC, Gallichio HA, Pye D, Walden NB (19771Application of immunofluorescence to the estab-lishment of an Encephalitozoon cuniculi-free rab-bit colony. Laboratory Animal Science 27, 204-9

Desplazes P, Mathis A, Baumgartner R, Tanner I,Weber R 119961Immunologic and molecular char-acterization of E. cuniculi-like microsporidia fromhumans and rabbits indicates that E. cuniculi is azoonotic parasite. Clinical Infectious Disease 22,557-9

Didier E5, Vossbrinck CR, Baker MD, Rogers LB,Bertucci DC, Shadduck JA (1995) Identification andcharacterization of three E. cuniculi strains. Para-Sitology 111, 411-21

Fedorko DP, Hijazi YM (19961 Application of mole-cular techniques to the diagnosis of microsporidialinfection. Emerging Infectious Diseases 2, 183-91

Greenstein G, Drozdowicz CK, Garcia FG, Lewis LL(19911 The incidence of Encephalitozoon cuniculiin a commercial barrier-maintained rabbit breedingcolony. Laboratory Animals 25,287-90

Hansen AK, Dagnaes-Hansen F, Mollegaard-HansenKE (19921 Correlation between megaloileitis andantibodies to Bacillus piliformis in laboratory ratcolonies. Laboratory Animal Science 42, 449-53




Hansen AK (1993) Statistical aspects of health mon-itoring of laboratory animal colonies. ScandinavianJournal of Laboratory Animal Science 20,11-14

Hansen AK, Skovgaard-Jensen HI (1995) Experience£tom sentinel health monitoring in units contain-ing rats and mice in experiments. ScandinavianJournal of Laboratory Animal Science 22, 1-9

Jacobson RH, Romatowski J (19961 Assessing thevalidity of serodiagnostic test results. Seminars inVeterinary Medical Surgery (Small Animal) 11,135-43

Kraft, V,Blanchet HM, Boot R, Deeny A, Hansen AK,Hem A, van Herck H, Kunstyr I, Needham JR,Nicklas W, Perrot A, Rehbinder C, Richard Y,deVroey G (1994) Recommendations for health mon-itoring of mouse, rat, hamster, guineapig and rabbitbreeding colonies. Laboratory Animals 28, 1-12

Pakes SP, Shadduck JA, Losen RG (1972) A diagnosticskin test for encephalitozoonosis (nosematosis) inrabbits. Laboratory Animal Science 22, 870-7

Pakes SP, Shadduck JA, Feldman DB, Moore JA (19841Comparison of tests for the diagnosis of sponta-neous encephalitozoonosis in rabbits. LaboratoryAnimal Science 34, 356-9

289

Pakes SP, Gerrity LW (1994) Protozoal diseases. In:The Biology of the Laboratory Rabbit (WeisbrothSH, Flatt RE, Kraus EL, eds). New York: AcademicPress, pp 205-29

Shadduck JA, Geroulo MJ (1979) A simple method forthe detection of antibodies to E. cuniculi in rabbits.Laboratory Animal Science 29, 330-4

Tyler JW, Cullor JS (1994) Titers, tests, and truisms:rational interpretation of diagnostic serologic test-ing. Journal of the American Veterinary Associa-tion 194, 1550-8

Waller T 119771The india-ink immunoreaction: amethod for the rapid diagnosis of encephalitozoo-nosis. Laboratory Animals 11, 93-7

Waller T, Morein B, Fabiansson E (1978) Humoralimmune response to infection with Encephalito-zoon cuniculi in rabbits. Laboratory Animals 12,145-8

Waller T 11988) Eradication of encephalitozoonosis inrabbit breeding colonies by carbon immunoassay.In: New Developments in Biosciences: TheirImplications for Laboratory Animal Science (Bey-nen AC, Solleveld HA, edsl. Dordrecht: MartinusNijhoff Publishers, pp 385-7




Documents

Comparison of Assays for Antibodies To