Diagnostic utility of CD4%:CD8low% T-lymphocyte ratio to differentiate feline immunodeficiency virus (FIV)-infected from FIV-vaccinated cats

Veterinary Microbiology 170 (2014) 197–205

Contents lists available at ScienceDirect

Veterinary Microbiology journal homepage: www.elsevier.com/locate/vetmic

Diagnostic utility of CD4%:CD8low% T-lymphocyte ratio to differentiate feline immunodeficiency virus (FIV)-infected from FIV-vaccinated cats Annette Litster a,*, Jui-Ming Lin b, Jamieson Nichols a, Hsin-Yi Weng c a

Department of Veterinary Clinical Sciences, School of Veterinary Medicine, Purdue University, 625 Harrison St., West Lafayette, IN 47907, USA Research and Development Department, IDEXX Laboratories, Westbrook, ME, USA c Department of Comparative Pathobiology, College of Veterinary Medicine, Purdue University, 785 Harrison Street, West Lafayette, IN 47907, USA b

A R T I C L E I N F O

A B S T R A C T

Article history: Received 11 June 2013 Received in revised form 17 January 2014 Accepted 24 January 2014

Antibody testing based on individual risk assessments is recommended to determine feline immunodeficiency virus (FIV) status, but ELISA and Western blot tests cannot distinguish between anti-FIV antibodies produced in response to natural infection and those produced in response to FIV vaccination. The aim of this cross-sectional study was to test the hypothesis that FIV-infected cats could be differentiated from FIV-vaccinated uninfected cats using lymphocyte subset results, specifically the CD4%:CD8low% Tlymphocyte ratio. Comparisons of the CD4%:CD8low% T-lymphocyte ratio were made among the following four groups: Group 1 – FIV-infected cats (n = 61; FIV-antibody positive by ELISA and FIV PCR positive); Group 2 – FIV-uninfected cats (n = 96; FIVantibody negative by ELISA); Group 3 – FIV-vaccinated uninfected cats (n = 31; FIVantibody negative by ELISA before being vaccinated against FIV, after which they tested FIV ELISA positive); and Group 4 – FIV-uninfected but under chronic/active antigenic stimulation (n = 16; FIV-antibody negative by ELISA; all had active clinical signs of either upper respiratory tract disease or gingival disease for  21 days). The median CD4%:CD8low% T-lymphocyte ratio was lower in Group 1 (1.39) than in each of the other three groups (Group 2 – 9.77, Group 3 – 9.72, Group 4 – 5.64; P < 0.05). The CD4%:CD8low% T-lymphocyte ratio was also the most effective discriminator between FIV-infected cats and the other three groups, and areas under ROC curves ranged from 0.91 (compared with Group 4) to 0.96 (compared with Group 3). CD4%:CD8low% shows promise as an effective test to differentiate between FIV-infected cats and FIV-vaccinated uninfected cats. ß 2014 Elsevier B.V. All rights reserved.

Keywords: Cat CD4 CD8low Diagnosis FIV Vaccine Retrovirus

1. Introduction Feline immunodeficiency virus (FIV) is a retrovirus of cats that primarily targets naı¨ve and activated CD4

* Corresponding author. Tel.: +1 765 418 3186. E-mail address: [email protected] (A. Litster). 0378-1135/$ – see front matter ß 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.vetmic.2014.01.014

T-lymphocyte subsets and monocytes, resulting in acute infection followed by persistent infection, similar to human immunodeficiency virus (HIV) infection in humans (Barlough et al., 1991; Hoffmann-Fezer et al., 1992; Bendinelli et al., 1995; Burkhard and Dean, 2003; Joshi et al., 2004). During the prolonged period of FIV latency in chronically infected cats, the CD4:CD8 T-cell ratio is expected to decrease because of both declining CD4 values and increased CD8 values (Tompkins et al., 1991; Murphy

198

A. Litster et al. / Veterinary Microbiology 170 (2014) 197–205

et al., 2012). In part, the increase in CD8 values in FIV infected cats is due to an expansion of a subset of Tlymphocytes, CD8low cells (Lehmann et al., 1992; Hofmann-Lehmann et al., 1995). The term CD8low refers to weak fluorescence of the b chain on the CD8ab glycoprotein and is seen in greater numbers in FIV infected cats (Shimojima et al., 1998). This expansion of CD8low cells can be seen as early as 4 weeks post-FIV infection and as late as 8 years post-infection (Willett et al., 1993; Shimojima et al., 1998). Although the exact role of the CD8low cells has not been completely defined, they show an up-regulation of MHC II expression compared to the control population (CD8high cells) found in FIV uninfected cats and could represent an activated T-cell subset reacting to the presence of FIV. CD8low cells also appear to be Lselectin negative, as in HIV-infected humans, and they contain a phenotype with strong anti-FIV suppressor activity, thought to be in response to FIV virus infection (Gebhard et al., 1999). A large study of client-owned and shelter-housed cats in the USA determined that the overall seroprevalence of FIV was 2.5%, and that the major risk factors were age, nonneutered gender status, clinical signs of ill health and an outdoor lifestyle (Levy et al., 2006). While the American Association of Feline Practitioners recommends antibody testing of all cats at appropriate intervals based on individual risk assessments to determine FIV status (Levy et al., 2008a), neither ELISA or Western blot tests can distinguish between anti-FIV antibodies produced in response to natural infection and those produced in response to FIV vaccination (Anderson and Tyrell, 2004; Crawford and Levy, 2007). Tests that detect virus, such as virus isolation or PCR can also be used for diagnosis. Virus isolation is considered the reference standard method, but it is not commercially available. There have been a number of published journal articles that report the diagnostic accuracy of FIV PCR testing (Anderson and Tyrell, 2004; Bienzle et al., 2004; Crawford et al., 2005; Wang et al., 2010; Morton et al., 2012; Ammersbach et al., 2013), but the diagnostic accuracy of assays varies widely and one study has suggested that FIV vaccination could interfere with the performance or interpretation of the assay (Crawford et al., 2005). There is a commercially available FIV PCR test for use by veterinary practitioners in the USA and other countries, but to our knowledge, there is just one conference presentation that reports the diagnostic accuracy of this particular test (Litster et al., 2012). Misdiagnosis of FIV infection in uninfected cats can have serious consequences, as it is a potential cause of inappropriate euthanasia (Crawford et al., 2005), especially in shelters where cat intake often exceeds the available resources and disposition decisions are sometimes based on the results of initial screening tests. Of course, accurate diagnosis is the starting point for effective case management and prevention of disease transmission, since FIVinfected cats generally remain free of clinical signs of infection for several years with informed husbandry and veterinary care (Hosie et al., 2009). The aim of this study was to investigate the diagnostic utility of combining ELISA and T-lymphocyte panel results to differentiate FIV-infected cats from FIV-vaccinates.

2. Materials and methods 2.1. Animals Cats were obtained for the study from shelters and households and all were examined by a veterinarian at enrollment. All laboratory tests were performed on aliquots of the same whole blood specimen from each cat. None of the cats had any history of vaccination against FIV prior to enrollment and they comprised four groups: Group 1 – FIV-infected group (n = 61)–These cats were enrolled in a separate ongoing prospective study and all were FIV antibody-positive1 and FIV PCR-positive.2 Twenty-three cats were privately owned in households of 1 or 2 cats/household, 28 cats were privately owned in households of >2 cats/household and 10 cats were shelterhoused (PAWS Chicago and Tree House Humane Society, Chicago). None of these cats had any history of vaccination against FIV. Group 2 – FIV-uninfected group (n = 96)–These cats were enrolled in the same ongoing prospective study as Group 1. All were FIV antibody-negative1 and FIV PCRnegative.2 Twenty-two cats were privately owned in households of 1 or 2 cats/household, 39 cats were privately owned in households of >2 cats/household and 35 cats were shelter-housed (PAWS Chicago and Tree House Humane Society, Chicago). Group 3 – FIV-vaccinated uninfected group (n = 31)– This group comprised adult cats that were FIV antibodynegative1 before they were fully vaccinated against FIV3 as recommended by the manufacturer. Whole blood was collected <12 months after the last vaccination was administered in each case. This group was comprised of 21 client-owned cats and 10 adult purpose-bred laboratory cats. The client-owned cats were kept in households of one or two cats each and the laboratory cats were individually cage-housed in a research facility.4 Blood was submitted to a reference laboratory5 for virus isolation after the vaccination course was complete and confirmed uninfected with FIV. All of these cats were clinically healthy on examination by a veterinarian at enrollment. Group 4 – FIV-uninfected chronic/active antigenic stimulation group (n = 16)–These cats were individually cage-housed in local shelters. All cats were FIV antibodynegative1 at the time blood was collected for the study and also at shelter intake some weeks previously. All but one of these cats had clinical signs of upper respiratory tract disease (ocular and/or nasal discharge, sneezing, coughing) on physical examination by a veterinarian at enrollment, while the remaining cat had signs of chronic dental disease (generalized gingivitis, purulent discharge at the gingival margin). This group was included so that results could be compared between chronically ill FIV-uninfected cats and

1 IDEXX SNAP1 FIV/FeLV ComboTM Test, IDEXX Laboratories, Westbrook, ME. 2 IDEXX FIV RealPCRTM Test, IDEXX Laboratories, West Sacramento, CA. 3 Fel-O-Vax1 FIV, Fort Dodge Animal Health, Fort Dodge, IA, now Pfizer Animal Health, New York, NY. 4 IDEXX Laboratories, Westbrook, ME 5 University of Glasgow Retrovirus Research Laboratory, Glasgow, UK.

A. Litster et al. / Veterinary Microbiology 170 (2014) 197–205

FIV-infected cats, to determine the effect of FIV infection, rather than chronic illness, on T-lymphocyte panel results. 2.2. Laboratory analysis Whole blood samples were collected by either jugular or saphenous venipuncture into EDTA tubes before transportation in insulated containers packed with ice bricks. The method is fully described elsewhere (Lin and Litster, 2013), but briefly, in a 12  75 mm tube, 100 mL fresh whole blood was mixed with antibodies against CD4,6 CD5,6 CD8,6 CD216 or CD61,7 conjugated with fluorescent dye. The antibodies were specifically made for use with feline blood specimens (Lin and Litster, 2013). These mixtures were incubated for 30 min at room temperature in dark conditions. After incubation, erythrocytes were lysed8 over approximately 15 min. The mixture was then centrifuged at 200  g for 5 min. The resultant supernatant was removed and the pellet was resuspended with 1 phosphate buffered saline9 (PBS)/1% fetal bovine serum10 (FBS). The resuspended cells were centrifuged at 200  g for 5 min. The resultant supernatant was removed and the cells were resuspended in 400 mL 1 PBS/1%FBS. The resuspended cells were examined and counted using a flow cytometer.11 Analysis of labeled cells was conducted using a flow cytometer.11 In this assay, forward scatter, side scatter and fluorescence were measured; the fluorescence channels used were fluorescein isothiocyanate, phycoerythrin and alexa fluoro 647.12 CD4 and CD8 bivariate plots were used to determine percentages of CD4 and CD8 T-lymphocytes. Ten thousand events were recorded in the fixed gate for lymphocytes (Fig. 1). The accuracy of the CD4% estimate was verified using a CD4 and CD5 co-label bivariate plot. For the CD8% estimate, accuracy was verified using a CD5 and CD8 co-label bi-variant plot. The methodology showed robust stability and precision over 3 days, yielding average day-to-day coefficients of variation of 2.15–9.35% for white blood cell (WBC) counts, lymphocyte counts, CD4 lymphocyte counts, CD8 lymphocyte counts and CD4:CD8, respectively (Lin and Litster, 2013). 2.3. Statistical analysis We first described and compared the distribution of leukocyte and flow cytometry variables in the four study groups. If the data were normally distributed (evaluated using Sharpiro–Wilk statistics), mean and standard deviation (SD) were reported and one-way ANOVA was used to compare the four group means. If the data were not normally distributed, median and interquartile range (IQR) were reported and the Kruskal–Wallis test was used to

6

AbD Serotech, Raleigh, NC. BD Biosciences, San Jose, CA. PharmLyse, BD Biosciences, San Jose, CA. 9 PBS, Mediatech Inc., Manassas, VA. 10 1% fetal bovine serum, SAFC, Sigma–Aldrich Corp., St. Louis, MO. 11 Accuri C6 flow cytometer, BD Accuri Cytometers, BD Biosciences, San Jose, CA. 12 Invitrogen, Carlsbad, CA. 7

199

compare the distribution among study groups. Pairwise comparisons with Bonferroni adjustment were performed if the P value from ANOVA or Kruskal–Wallis test was less than 0.05. Receiver operating characteristic (ROC) analysis was used to evaluate the ability of different T-lymphocyte variables to discriminate between FIV-infected cats (Group 1) and FIV-uninfected cats (Groups 2–4). Specifically, areas under curves (AUC) were reported and compared. Optimal cutoff values were determined using the Youden index, which is defined as the maximum of [sensitivity (1 specificity)] among all cutoff values. P values of < 0.05 were considered statistically significant. Commercially available software was used to draw graphs,13 and perform statistical analyses.14,15

3. Results 3.1. Animals Group 1 – FIV-infected group (n = 61) – There were 45 male neutered (MN) and 16 female spayed (FS) cats ranging in age from 2 to 10 years old (median age 5 years). At the time of specimen collection, 28 cats were clinically healthy on physical examination by a veterinarian. The remaining 33 cats had had one or more of the following abnormalities noted on physical examination: dental disease (n = 18), non-inflammatory alopecia (n = 5), skin wound (n = 1), thickened bowels on abdominal palpation (n = 5), cachexia (n = 2), unilateral or bilateral nasal discharge (n = 3), unilateral or bilateral ocular discharge (n = 3), systolic cardiac murmur on auscultation (n = 2), pustules on chin (n = 1), history of diabetes mellitus (n = 1), skin mass on chin (n = 1), bilateral or unilateral entropion (n = 2), active skin lesions diagnosed as ringworm (n = 1), active skin lesions diagnosed as flea allergy dermatitis (n = 2), skin mass (n = 1), and bilateral aural discharge (n = 1).The median time from first diagnosis of FIV by serology1 until specimen collection was 31 months (IQR 14.5–52.5 months). Group 2 – FIV-uninfected group (n = 96) – There were 64 MN and 32 FS cats ranging in age from 1.5 to 15 years old (median age 4 years). At the time of specimen collection, 80 cats were clinically healthy on physical examination by a veterinarian. The remaining 16 cats had one or more of the following abnormalities noted on physical examination: alopecia (n = 3), faucitis (n = 1), feline oral resorptive lesions (n = 3), malunion fracture (n = 1), mild dehydration (n = 1), pale mucous membranes (n = 1), bilateral aural discharge (n = 2), irregular heart rhythm with bradycardia (n = 1), unilateral ocular and nasal discharge (n = 1), subcutaneous mass near trachea (n = 1), skin wound on face (n = 1), loose canine teeth with associated periodontal disease (n = 1), inflamed gingivae that bled when touched (n = 1), serous unilateral ocular discharge (n = 1), and unilateral nasal discharge (n = 1).

8

13 GraphPad Prism 5.0d Macintosh version, Software MacKiev 1994– 2011 GraphPad Software Inc., San Diego, CA. 14 MedCalc version 12.2.1, MedCalc Software, Mariakerke, Belgium. 15 SPSS version 19 for Windows, SPSS Inc., Armonk, NY.

200

A. Litster et al. / Veterinary Microbiology 170 (2014) 197–205

Fig. 1. Light scatter plots of lymphocytes and large platelets and/or platelet aggregates and bivariate analysis plot of CD4 and CD8 T-lymphocytes. (a) Lymphocytes and large platelets and/or platelet aggregates overlap each other in the open square. (b) After excluding the CD61-labeled large platelets and/ or platelet aggregates, the lymphocytes were counted without platelet interference. (c) Open rectangles A and B contain CD8 T-lymphocytes; open rectangle B contains CD8low T-lymphocytes; and open rectangle C contains CD4 T-lymphocytes. FSC-A, forward side scatter; SSC-A, right angle scatter.

Group 3 – FIV-vaccinated uninfected group (n = 31)–The 21 client-owned cats had been fully vaccinated against FIV3 according to the manufacturer’s recommendations after pre-vaccination screening for FIV antibodies yielded negative results.1 After vaccination, all cats tested FIV antibodypositive1 and FIV PCR-negative.2 Blood from these cats was submitted to a reference laboratory5 after vaccination3 for virus isolation at the time of specimen collection and all were found to be uninfected with FIV. The laboratory cats (n = 10) were FIV antibody-negative1 on intake into a research facility.4 They were then vaccinated against FIV3 three times 2 weeks apart, as recommended by the manufacturer and housed without access to other cats. Specimen collection was performed in these cats 5 days after the third vaccination. In this group there were 10 MN, 11 FS, 4 male intact and 6 female intact cats and they ranged in age from 1 to 17.5 years old (median age 4.5 years). Group 4 – FIV-uninfected chronic/active antigenic stimulation group (n = 16) – The duration of clinical signs

was at least 3 weeks for all cats, and 15/16 cats showed signs of chronic upper respiratory tract disease (ocular and/or nasal discharge and/or sneezing). The remaining cat had severe gingivitis with mucopurulent discharge at the gingival margin. None of the cats in this group had access to FIV-positive cats at the time of sampling. In this group there were 7 MN, 9 FS cats and they ranged in age from 1 to 12 years old (median age 3 years). The ages of the cats at the time of specimen collection are shown in Fig. 2. There were no significant differences between the groups when a statistical comparison of age at the time of specimen collection was made. 3.2. Flow cytometry Based on the Sharpiro–Wilk statistics, T-lymphocyte measures were not normally distributed in all study groups. We therefore reported median and IQR and used Kruskal-Wallis tests to compare the distribution among

A. Litster et al. / Veterinary Microbiology 170 (2014) 197–205

201

20

Age (years)

15

10

5

4 up ro G

G

ro

up

3

2 up ro G

G

ro

up

1

0

Fig. 2. Box and whisker plots of the age of cats in each group (years) at the time of specimen collection. The central line of box–whisker plot is the median; the top and bottom of the box are the 75th and 25th centiles, respectively; the top and bottom whiskers are the 95th and 5th centiles, respectively.

study groups. CD4% for the FIV-infected group (Group 1; median 18.2, IQR 15.6–26.4) was significantly lower than for the other three FIV-uninfected groups (Fig. 3; Groups 2–4). The distribution of CD4% was similar among the FIVuninfected groups (Group 2: median 33.9, IQR 27.1–41.1; Group 3: median 34.4, IQR 25.1–42.8; Group 4: median 34.1, IQR 27.0–39.1).

Fig. 3. Box and whisker plots of percent of CD4 for four groups of cats. Groups 1–4 are FIV-infected group (n = 61), FIV-uninfected group (n = 96), FIV-vaccinated uninfected group (n = 31), and FIV-uninfected chronic/ active antigenic stimulation group (n = 16), respectively. The central line of box–whisker plot is the median; the top and bottom of the box are the 75th and 25th percentiles, respectively; the top and bottom whiskers are the maximum and minimum excluding outliers (circles) and extreme values (squares), respectively. Significant results from pairwise comparisons are presented by the horizontal lines connecting between the compared groups. ***P < 0.001.

Fig. 4. Box and whisker plots of percent of CD8low for four groups of cats. Groups 1–4 are FIV-infected group (n = 61), FIV-uninfected group (n = 96), FIV-vaccinated uninfected group (n = 31), and FIV-uninfected chronic/ active antigenic stimulation group (n = 16), respectively. The central line of box–whisker plot is the median; the top and bottom of the box are the 75th and 25th percentiles, respectively; the top and bottom whiskers are the maximum and minimum excluding outliers (circles) and extreme values (squares), respectively. Significant results from pairwise comparisons are presented by the horizontal lines connecting between the compared groups. *0.01  P < 0.05; ***P < 0.001.

In contrast to the CD4% results, the FIV-infected group (Group 1) had a higher median CD8low% (median 14.2, IQR 8.1–23.3) than the FIV-uninfected groups (Groups 2–4; Fig. 4). In addition, CD8low% was higher in the 33 FIVinfected cats that had at least one clinical abnormality identified on physical examination (median 15.4, IQR 11.5–23.5) than in the 28 clinically healthy FIV-infected cats (P = 0.039; median 8.9, IQR 6.5–19.2). The median CD8low% ranged from 2.9 in the FIV-vaccinated uninfected group (Group 3) to 5.1 in the FIV-uninfected chronic/active antigenic stimulation group (Group 4). The FIV-infected group (Group 1) also had a higher median CD8% (median 24.6, IQR 15.6–31.1) than the FIV-uninfected groups (Group 2–4) but this difference was not statistically significant (P = 0.119, Kruskal–Wallis test; Fig. 5). When using the ratio measures i.e. CD4%:CD8% and CD4%:CD8low%, the FIV-infected group (Group 1) yielded significantly lower results than the FIV-uninfected groups (Groups 2–4) and no difference was observed among the three FIV-uninfected groups (Groups 2–4; Figs. 6 and 7). The median CD4%:CD8% and CD4%:CD8low% for the FIVinfected group (Group 1) were 1.0 (IQR 0.55–1.23) and 1.39 (IQR 0.83–2.48), respectively. The median CD4%:CD8% was very similar among FIV-uninfected groups (Groups 2–4), ranging from 1.63 in the FIV-uninfected chronic/active antigenic stimulation group (Group 4) to 1.72 in the FIVvaccinated uninfected group (Group 3). The median CD4%:CD8low% ranged from 5.64 in the FIV-uninfected chronic/active antigenic stimulation group (Group 4) to 9.77 in the FIV-uninfected group (Group 2). Additionally, the median CD4%:CD8low% was lower in FIV-infected cats that had at least one clinical abnormality identified on physical examination (median 1.04, IQR 0.65–1.86) than in clinically healthy FIV-infected cats (median 1.97, IQR 1.10– 3.65; P = 0.005).

202

A. Litster et al. / Veterinary Microbiology 170 (2014) 197–205

Fig. 5. Box and whisker plots of percent of CD8 for four groups of cats. Groups 1–4 are FIV-infected group (n = 61), FIV-uninfected group (n = 96), FIV-vaccinated uninfected group (n = 31), and FIV-uninfected chronic/ active antigenic stimulation group (n = 16), respectively. The central line of box–whisker plot is the median; the top and bottom of the box are the 75th and 25th percentiles, respectively; the top and bottom whiskers are the maximum and minimum excluding outliers (circles) and extreme values (squares), respectively.

Fig. 7. Box and whisker plots of the ratio of percent of CD4 and percent of CD8 (CD4%:CD8%) for four groups of cats. Groups 1–4 are FIV-infected group (n = 61), FIV-uninfected group (n = 96), FIV-vaccinated uninfected group (n = 31), and FIV-uninfected chronic/active antigenic stimulation group (n = 16), respectively. The central line of box–whisker plot is the median; the top and bottom of the box are the 75th and 25th percentiles, respectively; the top and bottom whiskers are the maximum and minimum excluding outliers (circles) and extreme values (squares), respectively. Significant results from pairwise comparisons are presented by the horizontal lines connecting between the compared groups. *0.01  P < 0.05; ***P < 0.001.

and the FIV-vaccinated uninfected group (Group 3; P = 0.404; Table 2). The performance of CD4%:CD8low% was comparable to CD4% in distinguishing between the FIV-infected group (Group 1) and the FIV-uninfected chronic/active antigenic stimulation group (Group 4; P = 0.639; Table 3). Additional information is provided concerning different cutoff values and corresponding sensitivity and specificity for CD4%:CD8low% for distinguishing between the FIV-infected group (Group 1) and the FIV-vaccinated uninfected group (Group 3) in Table 4.

4. Discussion Fig. 6. Box and whisker plots of the ratio of percent of CD4 and percent of CD8low (CD4%:CD8low%) for four groups of cats. Groups 1–4 are FIVinfected group (n = 61), FIV-uninfected group (n = 96), FIV-vaccinated uninfected group (n = 31), and FIV-uninfected chronic/active antigenic stimulation group (n = 16), respectively. The central line of box–whisker plot is the median; the top and bottom of the box are the 75th and 25th percentiles, respectively; the top and bottom whiskers are the maximum and minimum excluding outliers (circles) and extreme values (squares), respectively. Significant results from pairwise comparisons are presented by the horizontal lines connecting between the compared groups. **0.001  P < 0.01; ***P < 0.001.

3.3. ROC analysis The results of ROC analyses are summarized in Tables 1–3. CD4%:CD8low% consistently outperformed the other Tlymphocyte variables in distinguishing between the FIVinfected group (Group 1) and the FIV-uninfected groups (Groups 2–4) with the following exceptions. Both CD4%:CD8low% and CD8low% performed equally well in distinguishing between the FIV-infected group (Group 1)

This study reports a new method for differentiating FIVinfected from FIV-vaccinated cats, which is a common problem wherever the FIV vaccine is used and can result in the euthanasia of uninfected healthy cats. A discriminatory ELISA method with 97.1% sensitivity and 100% specificity was reported in 2008 (Levy et al., 2008b), but this method has not been made available for commercial use at this time. While the sensitivity of the commercially available point-of-care FIV ELISA1,16 for detecting FIV-infected cats among cats not vaccinated against FIV is very high (100%; 95% CI 91–100%), it cannot reliably distinguish FIVinfected cats from FIV-vaccinated uninfected cats (specificity 0%; 95% CI 0–9%; Levy et al., 2004). Differentiating infected from vaccinated animals (DIVA) is a vital tool in feline practice and in animal shelters, as correct identification of FIV-infection status is essential for both individual

16

ME.

IDEXX SNAP1 Feline TripleTM Test, IDEXX Laboratories, Westbrook,

A. Litster et al. / Veterinary Microbiology 170 (2014) 197–205

203

Table 1 Receiver operating characteristic analysis of percent CD4 (CD4%), percent of CD8 (CD8%), percent of CD8low (CD8low%), ratio of percent of CD4 and percent of CD8 (CD4%:CD8%), and ratio of percent of CD4 and percent of CD8low (CD4%: CD8low%) for discriminating between Group 1 FIV-infected cats (n = 61) and Group 2 FIV-uninfected cats (n = 96). Flow cytometry variable

Area under curve (95% CIa)

Optimal cutoffb

Sensitivity (%)

Specificity (%)

CD4% CD8% CD8low% CD4%:CD8% CD4%: CD8low%

0.85 0.61 0.88 0.78 0.91

23.8 >27.5 >5.9 1.11 3.97

71 46 97 69 90

84 84 73 83 82

a b

(0.78–0.90) (0.52–0.68) (0.81–0.92) (0.71–0.84) (0.86–0.95)

95% confidence intervals. Selected based on the Youden index, which is defined as the maximum of [sensitivity

(1

specificity)].

Table 2 Receiver operating characteristic analysis of percent CD4 (CD4%), percent of CD8 (CD8%), percent of CD8low (CD8low%), ratio of percent of CD4 and percent of CD8 (CD4%:CD8%), and ratio of percent of CD4 and percent of CD8low (CD4%: CD8low%) for discriminating between Group 1 FIV-infected cats (n = 61) and Group 3 FIV-vaccinated uninfected cats (n = 31). Flow cytometry variable

Area under curve (95% CIa)

Optimal cutoffb

Sensitivity (%)

Specificity (%)

CD4% CD8% CD8low% CD4%:CD8% CD4%: CD8low%

0.84 0.60 0.94 0.77 0.95

28.1 >26.3 >5.9 1.21 5.92

80 49 97 75 98

71 87 87 77 84

a b

(0.75–0.91) (0.49–0.70) (0.88–0.98) (0.67–0.85) (0.89–0.99)

95% confidence intervals. selected based on the Youden index, which is defined as the maximum of [sensitivity

(1

specificity)].

Table 3 Receiver operating characteristic analysis of percent CD4 (CD4%), percent of CD8 (CD8%), percent of CD8low (CD8low%), ratio of percent of CD4 and percent of CD8 (CD4%:CD8%), and ratio of percent of CD4 and percent of CD8low (CD4%: CD8low%) for discriminating between Group 1 FIV-infected cats (n = 61) and Group 4 FIV-uninfected chronic/active antigenic stimulation cats (n = 16). Flow cytometry variable

Area under curve (95% CIa)

Optimal cutoffb

Sensitivity (%)

Specificity (%)

CD4% CD8% CD8low% CD4%:CD8% CD4%: CD8low%

0.89 0.52 0.84 0.75 0.91

22.3 >21.1 >5.2 1.15 2.42

67 56 97 72 75

100 63 56 81 94

a b

(0.79–0.95) (0.40–0.64) (0.74–0.91) (0.64–0.84) (0.82–0.96)

95% confidence intervals. Selected based on the Youden index, which is defined as the maximum of [sensitivity

animal care and population management (Crawford and Levy, 2007). When developing diagnostic strategies using tests that generate continuous variables, the choice of cutoff point determines the balance between test sensitivity and specificity. The choice of cutoff point is also determined by the individual clinical problem to be solved i.e. is it more

Table 4 Cutoff values of CD4%: CD8low% and their corresponding sensitivity and specificity for discriminating between Group 1 FIV-infected cats (n = 61) and Group 3 FIV-vaccinated uninfected cats (n = 31). Cutoff

Sensitivity

95% CIa

Specificity

95% CIa

3.06 3.58 4.82 5.92b 6.68

82 87 93 98 100

70–91 76–94 84–98 91–100 94–100

97 94 87 84 74

83–100 79–99 70–96 66–95 55–88

a

95% confidence intervals. Optimal cutoff value selected based on the Youden index, which is defined as the maximum of [sensitivity (1 specificity)]. b

(1

specificity)].

important to correctly rule out the possibility of FIVinfection, or to correctly identify FIV infected cats? This varies on an individual case basis, but in general terms, when managing an infectious disease in a shelter setting, test sensitivity is often prioritized over specificity to reduce the risk of exposing susceptible animals to disease. However, once a positive result has been obtained with an initial high sensitivity test, it is important to follow up on animals with positive results using a highly specific test, so that the possibility of infection can be ruled out in uninfected cats to avoid incorrect clinical decisions. In the FIV-infected group (n = 61) in this study, 28 (46%) cats were free of clinical abnormalities at the time of specimen collection and there was a wide range of timespans from first diagnosis of FIV by serology1 until specimen collection (median 31 months; IQR 14.5–52.5 months), leading us to conclude that this group represented a wide clinical spectrum of the course and effects of FIV infection. The inclusion of FIV-infected cats that did not have clinical signs of infection and/or were relatively recently diagnosed in the FIV-infected group, tested the

204

A. Litster et al. / Veterinary Microbiology 170 (2014) 197–205

sensitivity of the lymphocyte subset variables for their discriminatory ability, as it is known that changes in lymphocyte subsets become more pronounced with disease progression (Tompkins et al., 1991; Murphy et al., 2012). Interestingly, in the current study, CD8low% was statistically higher in FIV-infected cats that had clinical abnormalities on physical examination than in clinically healthy FIV-infected cats (P = 0.039). It has been shown that there can be age-related changes in feline lymphocyte subsets from late gestation until 3 months of age (Sellon et al., 1996; Bortnick et al., 1999). Furthermore, age-related changes in lymphocyte subsets were demonstrated in a study comparing adult cats (2–5 years old) and senior cats (aged 10–14 years), but these changes were primarily reported for absolute cell numbers rather than relative values (Campbell et al., 2004). However, the same study also reported lower CD4% in senior cats (Campbell et al., 2004), which could have contributed to the results for some senior cats in our study. However, there were no differences in age at the time of specimen collection when the four groups were compared statistically, so this was unlikely to have impacted the overall significance of the median comparisons between FIV infected and FIV uninfected cats. For the best performing T-lymphocyte variable in this study, CD4%:CD8low%, the optimal cutoff was chosen to give a 98% sensitivity and a 84% specificity to distinguish between FIV-infected and FIV-vaccinated uninfected cats (Table 2). The results suggest that CD4%:CD8low% could be used as a DIVA test after initial positive ELISA testing. In addition, one may use Table 4 to select different cutoff values based on their objectives. We would note that while sensitivity and specificity are not affected by prevalence of disease, the probabilities of using CD4%:CD8low to correctly differentiate between vaccinated and infected cats, which are more important in practice, depend in part on local FIV infection prevalence and the number of cats vaccinated against FIV. These posterior probabilities are also called positive predictive value and negative predictive value and their applications have been described elsewhere (Greiner and Gardner, 2000). An adjuvanted, inactivated whole-virus vaccine3 containing FIV subtype A (Petaluma) and subtype D (Shizuoka) has been licensed for use in many countries around the world, including the USA (Levy et al., 2004; Lecollinet and Richardson, 2008), but there have been conflicting reports of its efficacy in the published veterinary literature, especially concerning protection against subtype A (Huang et al., 2004; Kusuhara et al., 2005; Pu et al., 2005; Dunham et al., 2006). While little published data on the prevalence of FIV subtypes in the USA and Canada exist, subtypes A, B, C and F have been documented, with subtype B predominating in the USA (Weaver, 2010) and subtype A predominating in Canada (Reggeti and Bienzle, 2004). Accurate determination of FIV status before embarking on FIV vaccination is important and clients should be informed that their cats will have positive FIV ELISA results for an indeterminate but prolonged period after vaccination (Levy et al., 2008b). The point of care FIV ELISA tests currently licensed in the USA to detect anti-FIV antibodies in cats1,16 detect antibodies to the FIV core gag

proteins p24 and p15 which are antigenically similar across subtypes (Levy et al., 2004). It has been shown in multiple studies that the expansion of CD8low cells is specific to FIV infected cats in both the acute stage of infection and is persistent throughout the chronic symptomatic phase of disease (Willett et al., 1993; Bucci et al., 1998a; Shimojima et al., 1998; Gebhard et al., 1999). In experimentally infected cats, these cells have shown strong anti-FIV activity in invitro assays and their appearance correlates with a decrease in plasma- and cell-associated viremia, although the direct role of CD8low T-lymphocytes in decreasing viremia is not completely understood (Bucci et al., 1998a,b). In this study, FIV-infected cats with clinical abnormalities on physical examination had a higher CD8low% than healthy FIV-infected cats and this difference could possibly reflect a response to changes in viral load, health condition and/or FIV disease progression. However, viral load results were not available for direct comparison and given that these parameters have not been specifically studied in long-term FIV-infected cats, the exact cause for this discrepancy between FIV-infected cats of differing physical exam status remains unknown. While every effort was made to correctly determine the FIV infection status of all cats included in this study using the most accurate diagnostic methods available, since the available tests do not have perfect sensitivity and specificity, case definition remains a limitation of this study. In conclusion, while the ability of CD4%:CD8low% to differentiate between FIV-infected cats and FIV-vaccinated uninfected cats depends in part on local infection prevalence and the number of cats vaccinated against FIV, CD4%:CD8low% shows promise as an effective discriminatory test. Acknowledgements This study was supported, in part, by a grant from the Maddie’s Fund1. The Purdue Maddie’s Shelter Medicine Program is underwritten by a grant from Maddie’s Fund1, The Pet Rescue Foundation (www.maddiesfund.org), helping to fund the creation of a no-kill nation. We are grateful to Professor Margaret Hosie, Dr. Ayman Samman and Dr. Matthew Harris from the University of Glasgow Centre for Virus Research for performing the virus isolation work for this study. We would like to thank Dr. Gary D. Norsworthy and the staff at Alamo Feline Health Center, San Antonio, TX. We are also grateful to the staff and cats at PAWS Chicago and Tree House Humane Society, Chicago, IL; the Fitzhugh B. Crews FIV Cat Sanctuary, Jasper, GA; Almost Home Humane Society, Lafayette, IN; The Anti-Cruelty Society, Chicago, IL; and Jill Saucier, IDEXX Laboratories, Westbrook, ME for their assistance with data collection. References Ammersbach, M., Little, S., Bienzle, D., 2013. Preliminary evaluation of a quantitative polymerase chain reaction assay for diagnosis of feline immunodeficiency virus infection. J. Fel. Med. Surg. 15, 725–729. Anderson, P.R., Tyrell, P., 2004. Feline immunodeficiency virus diagnosis after vaccination. Anim. Health Res. Rev. 5, 327–330.

A. Litster et al. / Veterinary Microbiology 170 (2014) 197–205 Barlough, J.E., Ackley, C.D., George, J.W., Levy, N., Acevedo, R., Moore, P.F., Rideout, B.A., Cooper, M.D., Pedersen, N.C., 1991. Acquired immune dysfunction in cats with experimentally induced feline immunodeficiency virus infection: comparison of short-term and long-term infections. J. Acquir. Immune Defic. Syndr. 4, 219–227. Bendinelli, M., Pistello, M., Lombardi, S., Poli, A., Garzelli, C., Matteucci, D., Ceccherini-Nelli, L., Malvaldi, G., Tozzini, F., 1995. Feline immunodeficiency virus: an interesting model for AIDS studies and an important cat pathogen. Clin. Microbiol. Rev. 8, 87–112. Bienzle, D., Reggeti, F., Wen, X., Little, S., Hobson, J., Kruth, S., 2004. The variability of serological and molecular diagnosis of feline immunodeficiency virus infection. Can. Vet. J. 45, 753–757. Bortnick, S.J., Orandle, M.S., Papadi, G.P., Johnson, C.M., 1999. Lymphocyte subsets in neonatal and juvenile cats: comparison of blood and lymphoid tissues. Lab. Anim. Sci. 49, 395–400. Bucci, J.G., Gebhard, D.H., Childers, T.A., English, R.V., Tompkins, M.B., Tompkins, W.A., 1998a. The CD8+ cell phenotype mediating antiviral activity in feline immunodeficiency virus-infected cats is characterized by reduced surface expression of the CD8 b chain. J. Infect. Dis. 178, 968–977. Bucci, J.G., English, R.V., Jordan, H.L., Childers, T.A., Tompkins, M.B., Tompkins, W.A., 1998b. Mucosally transmitted feline immunodeficiency virus induces a CD8 antiviral response that correlates with reduction of cell-associated virus. J. Infect. Dis. 177, 18–25. Burkhard, M.J., Dean, G.A., 2003. Transmission and immunopathogenesis of FIV in cats as a model for HIV. Curr. HIV Res. 1, 15–29. Campbell, D.J., Rawlings, J.M., Koelsch, S., Wallace, J., Strain, J.J., Hannigan, B.M., 2004. Age-related differences in parameters of feline immune status. Vet. Immunol. Immunopathol. 100, 73–80. Crawford, P.C., Levy, J.K., 2007. New challenges for the diagnosis of feline immunodeficiency virus infection. Vet. Clin. North Am. Small Anim. Pract. 37, 335–350. Crawford, P.C., Slater, M.R., Levy, J.K., 2005. Accuracy of polymerase chain reaction assays for diagnosis of feline immunodeficiency virus infection in cats. J. Am. Vet. Med. Assoc. 226, 1503–1507. Dunham, S.P., Bruce, J., Mackay, S., Golder, M., Jarrett, O., Neil, J.C., 2006. Limited efficacy of an inactivated feline immunodeficiency virus vaccine. Vet. Rec. 158, 561–562. Gebhard, D.H., Dow, J.L., Childers, T.A., Alvelo, J.I., Tompkins, M.B., Tompkins, W.A., 1999. Progressive expansion of an L-selectin-negative CD8 cell with anti-feline immunodeficiency virus (FIV) suppressor function in the circulation of FIV-infected cats. J. Infect. Dis. 180, 1503– 1513. Greiner, M., Gardner, I.A., 2000. Application of diagnostic tests in veterinary epidemiologic studies. Prev. Vet. Med. 45, 43–59. Hoffmann-Fezer, G., Thum, J., Ackley, C., Herbold, M., Mysliwietz, J., Thefeld, S., Hartmann, K., Kraft, W., 1992. Decline in CD4 + cell numbers in cats with naturally acquired feline immunodeficiency virus infection. J. Virol. 66, 1484–1488. Hofmann-Lehmann, R., Holznagel, E., Aubertb, A., Ossent, P., Reinacher, M., Lutz, H., 1995. Recombinant FeLV vaccine: long-term protection and effect on course and outcome of FIV infection. Vet. Immunol. Immunopathol. 46, 127–137. Hosie, M.J., Addie, D., Bela´k, S., Boucraut-Baralon, C., Egberink, H., Frymus, T., Gruffydd-Jones, T., Hartmann, K., Lloret, A., Lutz, H., Marsilio, F., Pennisi, M.G., Radford, A.D., Thiry, E., Truyen, U., Horzinek, M.C., 2009. Feline immunodeficiency, ABCD guidelines on prevention and management. J. Feline Med. Surg. 11, 575–584. Huang, C., Conlee, D., Loop, J., Champ, D., Gill, M., Chu, H.J., 2004. Efficacy and safety of a feline immunodeficiency virus vaccine. Anim. Health Res. Rev. 5, 295–300. Joshi, A., Vahlenkamp, T.W., Garg, H., Tompkins, W.A., Tompkins, M.B., 2004. Preferential replication of FIV in activated CD4(+)CD25(+)T cells independent of cellular proliferation. Virology 321, 307–322. Kusuhara, H., Hohdatsu, T., Okumura, M., Sato, K., Suzuki, Y., Motokawa, K., Gemma, T., Watanabe, R., Huang, C., Arai, S., Koyama, H., 2005. Dual-subtype vaccine (Fel-O-Vax FIV) protects cats against contact challenge with heterologous subtype B FIV infected cats. Vet. Microbiol. 108, 155–165.

205

Lecollinet, S., Richardson, J., 2008. Vaccination against the feline immunodeficiency virus: the road not taken. Comp. Immunol. Microbiol. Infect. Dis. 31, 167–190. Lehmann, R., von Beust, B., Niederer, E., Condrau, M.A., Fierz, W., Aubert, A., Ackley, C.D., Cooper, M.D., Tompkins, M.B., Lutz, H., 1992. Immunization-induced decrease of the CD4+:CD8+ ratio in cats experimentally infected with feline immunodeficiency virus. Vet. Immunol. Immunopathol 35, 199–214. Levy, J.K., Crawford, P.C., Slater, M.R., 2004. Effect of vaccination against feline immunodeficiency virus on results of serologic testing in cats. J. Am. Vet. Med. Assoc. 225, 1558–1561. Levy, J., Crawford, C., Hartmann, K., Hofmann-Lehmann, R., Little, S., Sundahl, E., Thayer, V., 2008a. American Association of Feline Practitioners’ feline retrovirus management guidelines. J. Feline Med. Surg. 10, 300–316. Levy, J.K., Crawford, P.C., Kusuhara, H., Motokawa, K., Gemma, T., Watanabe, R., Arai, S., Bienzle, D., Hohdatsu, T., 2008b. Differentiation of feline immunodeficiency virus vaccination, infection, or vaccination and infection in cats. J. Vet. Intern. Med. 22, 330–334. Levy, J.K., Scott, H.M., Lachtara, J.L., Crawford, P.C., 2006. Seroprevalence of feline leukemia virus and feline immunodeficiency virus infection among cats in North America and risk factors for seropositivity. J. Am. Vet. Med. Assoc. 228, 371–376. Lin, J.M., Litster, A., 2013. Fluorescence flow cytometry methodology to exclude platelet aggregate interference when measuring feline CD4 and CD8 lymphocyte counts. Vet. J. 198, 275–278. Litster, A., Leutenegger, C., Estrada, M., Nichols, J., Hosie, M., Beczkowski, P., 2012. Diagnostic accuracy of a commercially available PCR test to detect feline immunodeficiency virus antigen. Proceedings of the 2nd Biennial Symposium of the International Society for Companion Animal Infectious Diseases, San Francisco, CA, Abstract #04 (abstract). Morton, J.M., McCoy, R.J., Kann, R.K., Gardner, I.A., Meers, J., 2012. Validation of real-time polymerase chain reaction tests for diagnosing feline immunodeficiency virus infection in domestic cats using Bayesian latent class models. Prev. Vet. Med. 104, 136–148. Murphy, B., Vapniarsky, N., Hillman, C., Castillo, D., McDonnel, S., Moore, P., Luciw, P.A., Sparger, E.E., 2012. FIV establishes a latent infection in feline peripheral blood CD4+ T lymphocytes in vivo during the asymptomatic phase of infection. Retrovirology 9, 12–17. Pu, R., Coleman, J., Coisman, J., Sato, E., Tanabe, T., Arai, M., Yamamoto, J.K., 2005. Dual-subtype FIV vaccine (Fel-O-Vax FIV) protection against a heterologous subtype B FIV isolate. J. Feline Med. Surg. 7, 65–70. Reggeti, F., Bienzle, D., 2004. Feline immunodeficiency virus subtypes A, B and C and intersubtype recombinants in Ontario, Canada. J. Gen. Virol. 85, 1843–1852. Sellon, R.K., Levy, J.K., Jordan, H.L., Gebhard, D.H., Tompkins, M.B., Tompkins, W.A., 1996. Changes in lymphocyte subsets with age in perinatal cats: late gestation through eight weeks. Vet. Immunol. Immunopathol. 53, 105–113. Shimojima, M., Takayuki, M., Kohmoto, M., Ikeda, Y., Nishimura, Y., Maeda, K., Tohya, Y., Mikami, T., 1998. Expansion of CD8a+b cells in cats infected with feline immunodeficiency virus. J. Gen. Virol. 79, 91–94. Tompkins, M.B., Nelson, P.D., English, R.V., Novotney, C., 1991. Early events in the immunopathogenesis of feline retrovirus infections. J. Am. Vet. Med. Assoc. 199, 1311–1315. Wang, C., Johnson, C.M., Ahluwalia, S.K., Chowdhury, E., Li, Y., Gao, D., Poudel, A., Rahman, K.S., Kaltenboeck, B., 2010. Dual-emission fluorescence resonance energy transfer (FRET) real-time PCR differentiates feline immunodeficiency virus subtypes and discriminates infected from vaccinated cats. J. Clin. Microbiol. 48, 1667–1672. Weaver, E.A., 2010. A detailed phylogenetic analysis of FIV in the United States. PLoS ONE 5, e12004. Willett, B.J., Hosie, M.J., Callanan, J.J., Ikeda, Y., Nishimura, Y., Maeda, K., Tohya, Y., Mikami, T., 1993. Infection with feline immunodeficiency virus is followed by the rapid expansion of a CD8+ lymphocyte subset. Immunology 78, 1–6.