Effect of chronic feline immunodeficiency infection, and efficacy of marbofloxacin treatment, on ‘Candidatus Mycoplasma haemominutum’ infection

Microbes and Infection 8 (2006) 653e661 www.elsevier.com/locate/micinf

Original article

Effect of chronic feline immunodeficiency infection, and efficacy of marbofloxacin treatment, on ‘Candidatus Mycoplasma haemominutum’ infection Se´verine Tasker*, Sarah M.A. Caney 1, Michael J. Day, Rachel S. Dean, Chris R. Helps, Toby G. Knowles, Philippa J.P. Lait, Mark D.G. Pinches, Timothy J. Gruffydd-Jones School of Clinical Veterinary Science, University of Bristol, Langford, Bristol, BS40 5DU, United Kingdom Received 14 April 2005; accepted 30 August 2005 Available online 17 January 2006

Abstract The purpose of this study was to investigate the effect of chronic feline immunodeficiency virus (FIV) infection, and efficacy of marbofloxacin treatment, on ‘Candidatus Mycoplasma haemominutum’ infection. Six cats chronically infected with FIV-Glasgow8 (group A) and six FIV-free cats (group B) were infected with ‘Candidatus M. haemominutum’ on day 0 by intravenous inoculation of blood. From day 0 to 105 post-infection (pi), blood samples were collected for ‘Candidatus M. haemominutum’ and FIV provirus quantitative real-time polymerase chain reaction (PCR) and haematological examination. Three of the six cats in each of the groups were randomly selected to receive marbofloxacin treatment (2 mg/kg PO SID) from day 49 to day 76 pi, with the remaining cats being untreated controls. Maximum ‘Candidatus M. haemominutum’ copy number was reached around day 30 pi. No overt cycling or marked variation in copy number was observed. No significant effect of FIV infection on ‘Candidatus M. haemominutum’ copy number kinetics or anaemia indices was found. No correlation was found between FIV provirus copy number and ‘Candidatus M. haemominutum’ copy number or haematological variables. Although marbofloxacin treatment was associated with a significant decrease in ‘Candidatus M. haemominutum’ copy number, the copy number plateaued during treatment, with no negative PCR results. Additionally, after termination of marbofloxacin treatment the copy numbers of the treated cats increased to reach levels similar to those of the untreated cats within 7e10 days. This study documents, for the first time, the infection kinetics and antibiotic responsiveness of ‘Candidatus M. haemominutum’ infection. Ó 2005 Elsevier SAS. All rights reserved. Keywords: Haemoplasma; Feline immunodeficiency virus; Quantitative real-time PCR

1. Introduction Studies have documented the existence of two distinct Haemobartonella felis species which have been named, in

* Corresponding author. School of Clinical Veterinary Science, University of Bristol, Langford House, Langford, Bristol, BS40 5DU, United Kingdom. Tel.: þ44 117 928 9558; fax: þ44 117 928 9559. E-mail address: [email protected] (S. Tasker). 1 Current address: Downland Veterinary Group, 71 Havant Road, Emsworth, Hants, PO10 7NZ, United Kingdom. 1286-4579/$ - see front matter Ó 2005 Elsevier SAS. All rights reserved. doi:10.1016/j.micinf.2005.08.015

line with their new classification as mycoplasmal species, as ‘Candidatus Mycoplasma haemominutum’ and Mycoplasma haemofelis [1e3]. These two species are collectively called the feline haemoplasmas. A recent report also documents the existence of a third pathogenic feline haemoplasma species in Switzerland [4] but further characterisation of this species is required. Experimental infection with M. haemofelis usually results in a severe, sometimes fatal, haemolytic anaemia [5e7]. However, experimental infection of cats with ‘Candidatus M. haemominutum’ does not usually result in significant clinical

654

S. Tasker et al. / Microbes and Infection 8 (2006) 653e661

signs or the development of anaemia [5,6]. However one experimental study [8] did find that mild or moderate anaemia was induced in seven of nine cats experimentally infected with ‘Candidatus M. haemominutum’, and ‘Candidatus M. haemominutum’-associated anaemia has been reported recently in a cat undergoing chemotherapy for lymphoma [9]. Further work on the pathogenicity of ‘Candidatus M. haemominutum’ isolates is required. Immunosuppression due to retrovirus infection [feline leukaemia virus (FeLV) and/or feline immunodeficiency virus (FIV)] can result in increased susceptibility to other infectious agents. An association between retroviral infection and ‘Candidatus M. haemominutum’ infection has been suggested. A recent study found that FeLV or FIV infected American feral cats were more likely to be haemoplasma infected than nonFeLV or non-FIV infected cats, respectively [10]. However, studies in UK feral cats, with a high prevalence of FIV infection, failed to show an association between FIV and haemoplasma infection (species not specified) [11], although cytology was used for diagnosis of haemoplasma infection, which is known to be unreliable compared to the polymerase chain reaction (PCR) [12]. Retroviral status may also influence the pathogenesis of feline haemoplasma infection. In one experimental study [8] involving the inoculation of ‘Candidatus M. haemominutum’, significantly lower haemoglobin values were reported in cats that were chronically infected with retroviruses (either FeLV infected or both FeLV and FIV infected) compared to those that were retrovirus-free. Another study reported more severe clinical signs (although statistical evaluation was not performed) due to haemoplasma infection (species not specified) in FIV infected cats compared to nonFIV infected cats [13]. In naturally infected cats in Israel [14], however, no significant difference was found in the severity of clinical signs or haematological parameters due to haemoplasma infection in cats infected with FIV compared to FIV-free cats. It is known that different FIV strains show varying pathogenicity [15], which may explain the differing results of these studies. However, further research is clearly required to define any effect of FIV infection on the pathogenesis of ‘Candidatus M. haemominutum’ infection. Studies evaluating the antibiotic responsiveness of ‘Candidatus M. haemominutum’ are limited. Two studies have evaluated small numbers (three or four cases) of ‘Candidatus M. haemominutum’-only infected cats and found that imidocarb dipropionate [16] and azithromycin [5] were not effective at clearing ‘Candidatus M. haemominutum’ infection. Although doxycycline and enrofloxacin have been evaluated as treatments for M. haemofelis infection [6,7,17,18], these antibiotics have not been evaluated for the treatment of ‘Candidatus M. haemominutum’ infection. The ability of a two week course of enrofloxacin to significantly reduce, but not consistently eliminate, M. haemofelis infection [17] suggests that a longer treatment period may be required to eliminate haemoplasma infection. Marbofloxacin, a fluoroquinolone antibiotic related to enrofloxacin, which is available throughout the world, has not yet been evaluated in the treatment of haemoplasmosis.

The inability to culture haemoplasmas has limited investigation of the kinetics of feline haemoplasma infection. The recent establishment of a sensitive and specific real-time PCR assay for quantification of feline haemoplasma copy number has allowed successful monitoring of haemoplasma copy number kinetics in vivo in a limited number of cats [19]. The aim of the current study was to use real-time PCR to evaluate the effect of chronic FIV infection on the kinetics of ‘Candidatus M. haemominutum’ infection. Additionally haematological parameters in both FIV infected and non-FIV infected cats experimentally infected with ‘Candidatus M. haemominutum’ were evaluated. A further aim of the study was to evaluate the efficacy of marbofloxacin in the treatment of ‘Candidatus M. haemominutum’ infection. 2. Materials and methods 2.1. Study design and protocol Twelve adult specific pathogen free (SPF)-derived, barriermaintained cats were used in this study. Six cats had been experimentally infected with FIV-Glasgow8 (isolate kindly donated by Professor O. Jarrett of the University of Glasgow) [20] for between 23 and 32 months and had remained clinically asymptomatic (group A). These comprised four male neutered and two female entire cats. The other six cats were known to be FIV-free (group B) and also comprised four male neutered and two female entire cats. Groups A and B were housed separately for the duration of the study. Blood samples (ethylenediaminetetraacetic acid [EDTA]anticoagulated) were collected on day
7 of the study for haematological examination (including reticulocyte count), packed cell volume determination (PCV), blood typing (RapidVet-H blood typing cards, DMS Laboratories Inc, New Jersey, US), ‘Candidatus M. haemominutum’ real-time PCR, FIV provirus real-time PCR and FIV antibody enzyme-linked immunosorbent assay testing (Idexx Laboratories, Wetherby, UK). Blood samples were also collected on day 0, just before inoculation with ‘Candidatus M. haemominutum’, for haematological examination, PCV, ‘Candidatus M. haemominutum’ real-time PCR, and FIV provirus realtime PCR. Experimental infection of all 12 cats with ‘Candidatus M. haemominutum’ was carried out by obtaining 24 ml of heparinised blood from an SPF-derived donor cat chronically infected with ‘Candidatus M. haemominutum’. Heparinised blood (2 ml) was given intravenously to all 12 cats within 10 min of collection from the donor. This rapid inoculation was achieved by pre-placing cephalic intravenous catheters in all recipient cats. The blood type of the donor cat was predetermined to be the same as all recipients. EDTA blood samples were collected from all 12 cats three times a week from day 0 to 105 post-infection (pi): twice weekly, 0.3 ml blood was collected for PCV and ‘Candidatus M. haemominutum’ real-time PCR, and once weekly, 0.7 ml blood was collected for PCV, haematological examination, ‘Candidatus M. haemominutum’ real-time PCR and FIV provirus real-time PCR. FIV provirus real-time PCR was

S. Tasker et al. / Microbes and Infection 8 (2006) 653e661

performed on all weekly samples collected from group A but only on those collected on day
7 and day 105 in group B. FIV antibody enzyme-linked immunosorbent assay testing was also performed on all cats on day 105 pi. On day 49 pi three of the six cats in each of groups A and B were randomly selected (Cats A2, A4 and A5, and Cats B2, B3 and B5) to receive marbofloxacin treatment (2 mg/kg PO SID) whilst the remaining three cats in each group remained as untreated controls. Treatment was given for a total of 28 days, until day 76 pi. 2.2. DNA extraction Briefly DNA was extracted from 100 ml of EDTA-anticoagulated blood using the DNeasy Tissue Kit (Qiagen, Crawley, UK) in accordance with the manufacturer’s instructions for whole blood, eluting into 200 ml of Buffer AE. DNA was stored at
20  C until evaluated in the PCR assays. 2.2.1. PCR assays ‘Candidatus M. haemominutum’ real-time PCR was carried out based on an assay previously described [19]. ‘Candidatus M. haemominutum’ real-time PCR was performed on all samples using 5 ml of DNA template per 25 ml reaction. Briefly, this assay uses primers (Hf forward 50 -ACGAAAGTCT GATGGAGCAATA-30 and Hf reverse 50 -ACGCCCAATAAA TCCGRATAAT-30 ) (Life Technologies, Paisley, Scotland) known to be specific for feline haemoplasmas together with a ‘Candidatus M. haemominutum’-specific Taqman probe (6-carboxyfluorescein [FAM]-AGCTTGATAGGAAATGAT TAAGCCTTGA-BHQ1) (Cruachem Ltd, Glasgow, Scotland). PCR reactions comprised 12.5 ml 2 Platinum Quantitative PCR Supermix-UDG (Invitrogen Ltd, Glasgow, UK), 260 nM of each primer, 150 nM ‘Candidatus M. haemominutum’ Taqman probe, 6 mM MgCl2 final concentration and 5 ml of template DNA, made up to a final volume of 25 ml with water. PCR was performed using an iCycler IQ system (Bio-Rad Laboratories Ltd, Hemel Hempstead, London) with an initial incubation at 50  C for 3 min, then 95  C for 2 min followed by 45 cycles of 95  C for 5 s and 58  C for 30 s. Fluorescence was detected at 530 nm at each annealing step. DNA samples from known ‘Candidatus M. haemominutum’-infected and non-infected cats, and water, were subjected to PCR as positive and negative controls, respectively. Quantification of ‘Candidatus M. haemominutum’ copy number in each PCR was derived by comparison to a standard curve generated by amplification of plasmids containing cloned 16S rDNA PCR products from ‘Candidatus M. haemominutum’, as described previously [19]. Briefly, plasmid copy number was determined spectrophotometrically and serial 10-fold dilutions were prepared for PCR following linearisation with EcoRI. These were subjected to ‘Candidatus M. haemominutum’ real-time PCR to enable generation of a standard curve to allow quantification of the ‘Candidatus M. haemominutum’ copy number in the unknown samples. FIV provirus real-time PCR was carried out using DNA extracted as described above. This assay comprises primers

655

(FIV 995 For 50 -TTAAGCCAGAAAGTACCCTAGAAG-30 and FIV 1133 Rev 50 -AAACACACTGGTCCTGATCC-30 ) (Life Technologies, Paisley, Scotland) and an FIV Taqman probe (FIV 1064 50 -FAM-TGCAACTCTTGGCAGAAGCT CTTACA-BHQ1-30 ) known to be specific for the Glasgow8 FIV gag gene. This assay has an efficiency of 102% and can detect down to 4 copies per PCR reaction. PCR reactions comprised 12.5 ml Hotstartaq Master mix (Qiagen, UK), 200 nM of each primer, 200 nM of FIV Taqman probe, 4.5 mM MgCl2 final concentration and 5 ml of template DNA, made up to a final volume of 25 ml with water. PCR was performed using an iCycler IQ system with an initial incubation at 95  C for 15 min, then 50 cycles of 95  C for 10 s and 60  C for 30 s. Fluorescence was detected at 530 nm at each annealing step. DNA samples from known FIV infected and non-infected cats, and water, were subjected to PCR as positive and negative controls, respectively. Quantification of FIV provirus copy number in each PCR was derived by comparison to a standard curve generated by PCR amplification of plasmids containing cloned FIV gag gene PCR products. As previously, plasmid copy number was determined spectrophotometrically. 2.3. Analysis of results The data were explored by plotting ‘Candidatus M. haemominutum’ and FIV provirus copy numbers and haematological data variables against time for each cat. Additionally plots of the mean values for all cats grouped into either FIV infected (group A) or non-FIV infected (group B), or into marbofloxacin-treated or controls, were also examined. To evaluate for any effect of FIV infection, during the pretreatment period (days 1e49 pi), ‘Candidatus M. haemominutum’ copy number and haematological data were compared between groups A and B using the ManneWhitney U test. The ‘Candidatus M. haemominutum’ copy number values were logged to allow values to be more biologically meaningful in view of the great variation in copy numbers recorded over time. The average logged ‘Candidatus M. haemominutum’ copy number for each cat over the pre-treatment time period was used in the analysis. For the haematological variables [PCV, haemoglobin (Hb), red blood cell count (RBC), mean cell volume (MCV), mean corpuscular haemoglobin concentration (MCHC), white blood cell count (WBC), neutrophil count (PMN), lymphocyte count, monocyte count, eosinophil count, platelet count, aggregate reticulocyte count and punctate reticulocyte count] summations of the values for each cat over the pre-treatment time period were used in a Manne Whitney U test analysis. The test for a difference between groups A and B was not significant, thus the groups were amalgamated for the test of treatment. To evaluate for any effect of marbofloxacin treatment, ‘Candidatus M. haemominutum’ copy number and haematological data recorded during the treatment period (days 50e76 pi) were compared between the treated and untreated cats as described above using the ManneWhitney U test. Exploration of the data showed an apparent fall in PCV from day 0 to 23 pi in both groups A and B following

S. Tasker et al. / Microbes and Infection 8 (2006) 653e661

656

Comparison of ‘Candidatus M. haemominutum’ copy numbers between days 51 and 77 pi inclusive (during the marbofloxacin treatment period) found a significant difference between the treated and untreated control cats (P ¼ 0.002). This is consistent with marbofloxacin treatment having a significant effect on ‘Candidatus M. haemominutum’ copy number. Figs. 1 and 2 clearly depict lower ‘Candidatus M. haemominutum’ copy numbers in the treated cats compared to the untreated control cats in both groups A and B. Figs. 1 and 2 also demonstrate an apparent plateau effect of marbofloxacin treatment, in that the ‘Candidatus M. haemominutum’ copy number in treated cats appeared to level out during the treatment period. Between days 77 and 104 pi, after the marbofloxacin antibiotic treatment period, the ‘Candidatus M. haemominutum’ copy number rose again in the treated cats to reach levels similar to those of the untreated cats within 7e10 days of stopping marbofloxacin treatment (Figs. 1 and 2).

‘Candidatus M. haemominutum’ inoculation. To evaluate whether ‘Candidatus M. haemominutum’ infection was associated with a change in PCV, a repeated measures one-way analysis of variance (ANOVA) test was performed for all 12 cats from day 0 to 23 pi. Note that this test was applied retrospectively, after an inspection of the data, and so the results can be viewed as a guide only. Statistical analysis was performed using Statistical Package for Social Scientists version 11.5. For all statistical analysis, significance was taken as a P value of <0.05. Correction factors were not applied because of the small sample sizes used in this study and the need to recognise any potentially important differences. 3. Results All cats were ‘Candidatus M. haemominutum’ PCR negative on days
7 and 0. All cats were blood type A. On days
7 and 105 pi all cats in group A were FIV antibody positive whilst all cats in group B were FIV antibody negative.

3.2. Haematological variables Figs. 3 and 4 show PCV variation over time for all cats in groups A and B, respectively. Comparison of haematological variables between days 2 and 49 pi inclusive, found no significant difference between groups A and B for PCV (P ¼ 0.59), Hb (P ¼ 0.09), RBC (P ¼ 0.31), MCV (P ¼ 0.82), MCHC (P ¼ 0.06), WBC (P ¼ 0.93), neutrophil count (P ¼ 1.00), lymphocyte count (P ¼ 0.82), monocyte count (P ¼ 0.82), eosinophil count (P ¼ 1.00), platelet count (P ¼ 0.82), aggregate reticulocyte count (P ¼ 0.47) and punctate reticulocyte count (P ¼ 1.00), indicating no significant effect of FIV infection on these parameters. Significance was, however, reached for the basophil count, with FIV infection being associated with significantly lower basophil counts (P ¼ 0.04). Although correction factors

3.1. ‘Candidatus M. haemominutum’ copy number Figs. 1 and 2 show ‘Candidatus M. haemominutum’ copy number variation over time for all cats in groups A and B, respectively. Comparison of ‘Candidatus M. haemominutum’ copy numbers between days 2 and 49 pi inclusive (before the start of marbofloxacin treatment) revealed no significant difference between groups A and B (P ¼ 0.18). This is consistent with concurrent FIV infection having no significant effect on ‘Candidatus M. haemominutum’ copy number. Indeed Figs. 1 and 2 do not show any striking difference in ‘Candidatus M. haemominutum’ copy number between groups A and B.

Day 0 - 48 pi

Day 49 - 76 pi

Day 77 - 104 pi

‘Candidatus M. haemominutum’ copy no. per PCR

Marbofloxacin treatment period Cats A2, A4 & A5 treated Cats A1, A3 & A6 untreated controls

100000000 10000000 1000000

CAT A1

100000 CAT A2

10000

CAT A3

1000

CAT A4

100

CAT A5

10

CAT A6

1.0 0.1 98

103

93

89

84

79

75

70

65

61

56

51

47

42

37

33

28

23

19

9

14

5

0

negative

Day post-infection (pi) Fig. 1. ‘Candidatus M. haemominutum’ copy number per PCR during study for group A cats. All group A cats were FIV infected. Grey-shaded area indicates marbofloxacin treatment period. Between days 49 and 76 pi inclusive Cats A2, A4 and A5 were treated with marbofloxacin whilst the remaining cats were untreated controls.

S. Tasker et al. / Microbes and Infection 8 (2006) 653e661 Day 0 - 48 pi

Day 49 - 76 pi

657

Day 77 - 104 pi

‘Candidatus M. haemominutum’ copy no. per PCR

Marbofloxacin treatment period Cats B2, B3 & B5 treated Cats B1, B4 & B6 untreated controls

100000000 10000000 1000000 100000

CAT B1 B1

10000

CAT B2 B2

1000

B3 CAT B3

100

B4 CAT B4

10

CAT B5 B5

1.0

CAT B6 B6

0.1 98

103

93

89

84

79

75

70

65

61

56

51

47

42

37

33

28

23

19

9

14

5

0

negative

Day post-infection (pi) Fig. 2. ‘Candidatus M. haemominutum’ copy number per PCR during study for group B cats. All group B cats were non-FIV infected. Grey-shaded area indicates marbofloxacin treatment period. Between days 49 and 76 pi inclusive Cats B2, B3 and B5 were treated with marbofloxacin whilst the remaining cats were untreated controls.

were not routinely applied, Bonferroni’s adjusted P value for comparison of the haematological variables between FIV infected and non-FIV infected cats was P < 0.0036, below the P value reported for the basophil count. Comparison of haematological variables between days 51 and 77 pi inclusive, found no significant difference between the treated and untreated control cats for PCV (P ¼ 0.56), Hb (P ¼ 0.82), RBC (P ¼ 0.49), MCV (P ¼ 0.15), MCHC (P ¼ 0.39), WBC (P ¼ 0.07), neutrophil count (P ¼ 0.70), monocyte count (P ¼ 0.24), eosinophil count (P ¼ 0.24), basophil count (P ¼ 0.55), platelet count (P ¼ 0.24), punctate reticulocyte count (P ¼ 0.06) and aggregate reticulocyte count (P ¼ 1.00), indicating no significant effect of marbofloxacin treatment on these parameters. Significance was, however, reached for the lymphocyte count, with marbofloxacin

treatment being associated with significantly higher lymphocyte counts (P ¼ 0.04). However, Bonferroni’s adjusted P value for comparison of the haematological variables between treated and untreated control cats was P < 0.0036, below the P value reported for the lymphocyte count. A repeated measures one-way analysis of variance test, performed on the PCV values for all 12 cats from day 0 to 23 pi, showed a significant decrease in PCV over time (P < 0.001). 3.3. FIV provirus copy number All six cats in group B were negative for FIV provirus DNA at all time points (days
7 and 105 pi) evaluated. The six cats in group A generated positive results for FIV provirus DNA, although Cat A2 was negative at eight of the 17 time points

Marbofloxacin treatment period Cats A2, A4 & A5 treated Cats A1, A3 & A6 untreated controls

70 60

CAT A1

PCV (%)

50 CAT A2 CAT A3

40

CAT A4

30

CAT A5 CAT A6

20

103

98

93

89

84

79

75

70

65

61

56

51

47

42

37

33

28

23

19

14

9

5

0

10

Day post-infection Fig. 3. PCV during study for group A cats. All group A cats were FIV infected. Grey-shaded area indicates marbofloxacin treatment period. Between days 49 and 76 pi inclusive Cats A2, A4 and A5 were treated with marbofloxacin whilst the remaining cats were untreated controls.

S. Tasker et al. / Microbes and Infection 8 (2006) 653e661

658

Marbofloxacin treatment period Cats B2, B3 & B5 treated Cats B1, B4 & B6 untreated controls

70 60

CAT B1

PCV (%)

50

CAT B2 CAT B3

40

CAT B4

30

CAT B5 CAT B6

20

98

103

93

84

89

75

79

65

70

61

56

47

51

42

37

33

28

19

23

9

14

5

0

10

Day post-infection Fig. 4. PCV during study for group B cats. All group B cats were non-FIV infected. Grey-shaded area indicates marbofloxacin treatment period. Between days 49 and 76 pi inclusive Cats B2, B3 and B5 were treated with marbofloxacin whilst the remaining cats were untreated controls.

reaching a maximum at around days 28e33 pi for both FIV and non-FIV infected cats. No evidence of an effect of chronic FIV infection on the kinetics of ‘Candidatus M. haemominutum’ infection was found. Similarly, between days 2 and 49 pi, no significant effect of FIV infection was seen on the majority of haematological variables following ‘Candidatus M. haemominutum’ infection, including indices of anaemia such as PCV, haemoglobin and RBC. A previous study [8] found that the inoculation of ‘Candidatus M. haemominutum’ caused significantly lower haemoglobin values in both FeLV infected and dual FeLV and FIV infected cats compared to those that were retrovirus-free. The current study is the first to evaluate ‘Candidatus M. haemominutum’ infection in FIV only infected cats but no effect of chronic FIV infection was seen. The lack of an effect of FIV infection on the ‘Candidatus M. haemominutum’ copy number and clinical syndrome seen in this study may be a manifestation of the stage and strain of FIV infection as well as the stage of ‘Candidatus M. haemominutum’ infection involved. The cats in the current study were

evaluated whilst Cat A6 was negative at one of the 17 time points evaluated. Fig. 5 shows the variation in FIV provirus copy number over time for the group A cats. Exploration of the data did not reveal any correlation between FIV provirus DNA copy number and ‘Candidatus M. haemominutum’ copy number or haematological variables. 4. Discussion

1000

100 CAT A1 CAT A2

10

CAT A3 CAT A4 CAT A5

1

CAT A6

105

98

91

84

77

70

63

56

49

42

35

28

21

14

7

0

negative -7

FIV provirus copy no. per µl blood

The current study is the first documenting the kinetics of ‘Candidatus M. haemominutum’ infection in both non-FIV infected and chronically FIV infected cats. Values on the y axis of Figs. 1 and 2 represent measured copy number per PCR which approximates to organism number per 2.5 ml of blood, assuming 100% efficiency of the DNA extraction process and PCR, and the presence of one copy of the 16S rRNA gene in the ‘Candidatus M. haemominutum’ genome. Figs. 1 and 2 show a marked increase in ‘Candidatus M. haemominutum’ copy number with time in the initial pi period,

Day post ‘Candidatus M. haemominutum’ infection Fig. 5. FIV provirus DNA copy numbers per ml of blood (calculated from the FIV provirus DNA copy numbers per PCR, assuming a 100% efficiency of the DNA extraction procedure and taking into account the 102% efficiency of the PCR) during study for group A cats. All group A cats were FIV infected.

S. Tasker et al. / Microbes and Infection 8 (2006) 653e661

all chronically infected with a FIV-Glasgow8 isolate, a reasonably virulent FIV isolate [15,20] but were clinically normal at the start of the study. It may be that a different outcome would have resulted if symptomatic FIV infected cats, or cats recently infected with FIV, had been inoculated with ‘Candidatus M. haemominutum’. Fig. 5 shows the FIV provirus copy number variation over time in the group A cats. Values on the y axis of Fig. 5 represent measured copy number per ml of blood, assuming close to 100% efficiency of the DNA extraction process and PCR. Although four of the six cats in group A remained positive for FIV provirus throughout the study, Cats A2 and A6 were FIV provirus negative on several occasions. These negative results may represent time points at which the FIV provirus concentration in the blood was below the sensitivity of the PCR assay (equivalent to 1.6 proviral copies per ml of blood) and are unlikely to indicate clearance of FIV infection. Indeed both Cats A2 and A6 did appear to have low overall FIV provirus copy numbers (usually less than 10 per ml of blood) throughout the study. Both of these cats were FIV antibody positive at days
7 and 105, indicative of persistent infection, although at these time points PCR also generated positive results. Exploration of the data did not reveal any evidence of correlation between FIV provirus DNA copy number and ‘Candidatus M. haemominutum’ copy number or haematological variables. This study documented a significant decrease in PCV between days 0 and 23 pi following ‘Candidatus M. haemominutum’ infection. Statistical analysis was performed over this time period due to the progressive fall in PCV seen between days 0 and 23 pi in both groups A and B. Evaluation of the pathogenicity of ‘Candidatus M. haemominutum’ per se was not a primary aim of this study (this would have necessitated inclusion of an uninfected control group), but it was felt that the apparent fall in PCV warranted further analysis. Throughout the study minimal volume blood samples were taken (1.3 ml/cat/week maximum) to avoid blood loss making a significant contribution to any haematological changes seen. Although a significant fall in PCV was seen, only two of the six cats in each of groups A and B had their PCV drop below the reference range (25e45%). The fall in PCV is nevertheless noteworthy, and suggests that ‘Candidatus M. haemominutum’ infection does induce some degree of RBC destruction. The fall in PCV was associated with a rise in aggregate reticulocyte count (data not shown) in both groups A (from zero on day 0 to a maximum of 33  109/l on day 35 pi) and B (from zero on day 0 to a maximum of 32  109/l on day 28 pi). Although these reticulocyte counts are not high enough to indicate a regenerative response, their rise and fall during the course of ‘Candidatus M. haemominutum’ infection is consistent with the presence of a degree of haemolysis. It is interesting to note, however, that the maximal fall in PCV occurred slightly sooner than the maximal ‘Candidatus M. haemominutum’ copy number. Despite the donor and recipient cats all being blood type A, it cannot be ruled out, however, that an incompatible transfusion reaction caused the decrease in PCV seen, rather than it being due to ‘Candidatus M.

659

haemominutum’ infection alone. Interestingly previous study has also documented a fall in Hb following ‘Candidatus M. haemominutum’ infection [8]. Although ‘Candidatus M. haemominutum’ appears to be less pathogenic than M. haemofelis, these results suggest that this organism may have some limited virulence. Two week courses of doxycycline or enrofloxacin have been shown to be effective in the treatment of M. haemofelis infection, although elimination of infection does not consistently occur [17]. However, doxycycline has been associated with oesophageal stricture formation [21e23], and enrofloxacin has been associated with diffuse retinal degeneration and acute blindness in cats [24]. We therefore chose to evaluate marbofloxacin for ‘Candidatus M. haemominutum’ treatment in the current study. A four week treatment course was used in order to facilitate elimination of infection and this resulted in a significant decrease in ‘Candidatus M. haemominutum’ copy number between days 51 and 77 pi inclusive. This suggests that marbofloxacin is an effective antibiotic for the treatment of ‘Candidatus M. haemominutum’ infection, although the fall in ‘Candidatus M. haemominutum’ copy number was not as rapid or dramatic as that observed in studies monitoring M. haemofelis copy number in response to doxycycline or enrofloxacin treatment [17]. A recent case report documented the successful treatment of ‘Candidatus M. haemominutum’ infection with doxycycline, in a cat undergoing chemotherapy for lymphoma [9], although the duration of treatment was not clear since owner compliance was an issue. In this case report the cat became PCR negative during doxycycline treatment, which is in contrast to the results of the current study in which PCR results remained highly positive throughout marbofloxacin treatment. Elimination of infection should be the aim of treatment for ‘Candidatus M. haemominutum’ infection since it can be pathogenic and clearance would prevent any subsequent recrudescence of disease and/ or transmission of infection to other cats. However, direct comparison of studies evaluating clearance of infection with treatment is difficult as the sensitivity of the PCR assay used in such studies will influence the results obtained. The sensitivity of the PCR assay used by De Lorimier and Messick [9], which documented negative PCR results, is not known whereas the sensitivity of the assay used in the current study is known to be 3.6 copies of ‘Candidatus M. haemominutum’ DNA per PCR [19]. No significant difference between the treated and untreated control cats was found for the majority of the haematological variables. The lack of a significant effect of marbofloxacin on RBC indices is not surprising considering that ‘Candidatus M. haemominutum’ infection was not associated with marked haematological changes at the time marbofloxacin treatment was started; although a fall in PCV was found with ‘Candidatus M. haemominutum’ infection between days 2 and 23 pi, Figs. 3 and 4 show that there was a tendency for the PCV to increase again between days 24 and 47 pi, just before marbofloxacin treatment was started. It may be that a significant effect of marbofloxacin would have been found if treatment had been initiated at the point of maximal fall in PCV

660

S. Tasker et al. / Microbes and Infection 8 (2006) 653e661

(i.e. day 23 pi) but the aim of this study was to evaluate the efficacy of marbofloxacin in eliminating infection rather than treating a reduction in PCV. In the current study all cats in each of group A and group B were housed together. Group housing of cats is commonly employed in such studies [5,17,18]. The mechanisms of transmission of haemoplasma infection in cats have yet to be defined. It is known that infected cats can be housed long term with noninfected cats with no evidence of horizontal transmission of infection [25]. Fleas have been implicated in transmission of ‘Candidatus M. haemominutum’ infection [26,27] but the cats in the current study were known to be ectoparasite free. Although within group transmission of ‘Candidatus M. haemominutum’ infection cannot be ruled out completely in the current study, the authors feel that reinfection is unlikely in view of the ‘Candidatus M. haemominutum’ copy number kinetics recorded during the study. For example, a significant effect of marbofloxacin treatment in lowering ‘Candidatus M. haemominutum’ copy number was observed despite the treated and untreated control cats in each group being housed together. The current study failed to demonstrate a significant effect of chronic FIV infection on ‘Candidatus M. haemominutum’ copy number kinetics or haematological changes due to ‘Candidatus M. haemominutum’ infection. Although a four week course of marbofloxacin treatment was associated with a significant decrease in ‘Candidatus M. haemominutum’ copy number, clearance of infection was not evident. Studies evaluating longer treatment courses with alternative antibiotics for ‘Candidatus M. haemominutum’ infection are required. Acknowledgements

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12] [13]

[14]

[15]

The authors wish to thank Simon Cue, Elsa Crawford, Kathy Egan, Ross Harley, Kostas Papasouliotis and Suzanne Rudd for their help with this study. This study was funded by a University of Bristol Research Grant. During this study Se´verine Tasker held a position funded by Waltham, Sarah Caney held a position funded by the Feline Advisory Bureau and Rachel Dean held a position funded by Fort Dodge Animal Health.

[17]

References

[19]

[1] H. Neimark, K.E. Johansson, Y. Rikihisa, J.G. Tully, Proposal to transfer some members of the genera Haemobartonella and Eperythrozoon to the genus Mycoplasma with descriptions of ‘Candidatus Mycoplasma haemofelis’, ‘Candidatus Mycoplasma haemomuris’, ‘Candidatus Mycoplasma haemosuis’ and ‘Candidatus Mycoplasma wenyonii’, Int. J. Syst. Evol. Microbiol. 51 (2001) 891e899. [2] J.E. Foley, N.C. Pedersen, ‘Candidatus Mycoplasma haemominutum’, a low-virulence epierythrocytic parasite of cats, Int. J. Syst. Evol. Microbiol. 51 (2001) 815e817. [3] H. Neimark, K.E. Johansson, Y. Rikihisa, J.G. Tully, Revision of haemotrophic Mycoplasma species names, Int. J. Syst. Evol. Microbiol. 52 (2002) 683. [4] B. Willi, F.S. Boretti, V. Cattori, S. Tasker, M.L. Meli, C. Reusch, H. Lutz, R. Hofmann-Lehmann, Identification, molecular characterisation and experimental transmission of a new hemoplasma isolate from

[16]

[18]

[20]

[21]

[22]

[23]

a cat with hemolytic anaemia in Switzerland, J. Clin. Microbiol. 43 (2005) 2581e2585. D.S. Westfall, W.A. Jensen, W.J. Reagan, S.V. Radecki, M.R. Lappin, Inoculation of two genotypes of Haemobartonella felis (California and Ohio variants) to induce infection in cats and the response to treatment with azithromycin, Am. J. Vet. Res. 62 (2001) 687e691. J.E. Foley, S. Harrus, A. Poland, B. Chomel, N.C. Pedersen, Molecular, clinical, and pathologic comparison of two distinct strains of Haemobartonella felis in domestic cats, Am. J. Vet. Res. 59 (1998) 1581e1588. L.M. Berent, J.B. Messick, S.K. Cooper, Detection of Haemobartonella felis in cats with experimentally induced acute and chronic infections, using a polymerase chain reaction assay, Am. J. Vet. Res. 59 (1998) 1215e1220. J.W. George, B.A. Rideout, S.M. Griffey, N.C. Pedersen, Effect of preexisting FeLV infection or FeLV and feline immunodeficiency virus coinfection on pathogenicity of the small variant of Haemobartonella felis in cats, Am. J. Vet. Res. 63 (2002) 1172e1178. L.P. De Lorimier, J.B. Messick, Anemia associated with ‘Candidatus Mycoplasma haemominutum’ in a feline leukemia virus-negative cat with lymphoma, J. Am. Anim. Hosp. Assoc. 40 (2004) 423e427. B.J. Luria, J.K. Levy, M.R. Lappin, E.B. Breitschwerdt, A.M. Legendre, J.A. Hernandez, S.P. Gorman, I.T. Lee, Prevalence of infectious diseases in feral cats in Northern Florida, J. Feline Med. Surg. 6 (2004) 287e296. N. Yamaguchi, D.W. Macdonald, W.C. Passanisi, D.A. Harbour, C.D. Hopper, Parasite prevalence in free-ranging farm cats, Felis silvestris catus, Epidemiol. Infect. 116 (1996) 217e223. S. Tasker, M.R. Lappin, Haemobartonella felis: recent developments in diagnosis and treatment, J. Feline Med. Surg. 4 (2002) 3e11. G.H. Reubel, G.A. Dean, J.W. George, J.E. Barlough, N.C. Pedersen, Effects of incidental infections and immune activation on disease progression in experimentally feline immunodeficiency virus-infected cats, J. Acquir. Immune. Defic. Syndr. 7 (1994) 1003e1015. S. Harrus, E. Klement, I. Aroch, T. Stein, H. Bark, E. Lavy, M. MazakiTovi, G. Baneth, Retrospective study of 46 cases of feline haemobartonellosis in Israel and their relationships with FeLV and FIV infections, Vet. Rec. 151 (2002) 82e85. M.J. Hosie, T. Dunsford, D. Klein, B.J. Willett, C. Cannon, R. Osborne, J. Macdonald, N. Spibey, N. Mackay, O. Jarrett, J.C. Neil, Vaccination with inactivated virus but not viral DNA reduces virus load following challenge with a heterologous and virulent isolate of feline immunodeficiency virus, J. Virol. 74 (2000) 9403e9411. M.R. Lappin, M. Brewer, S. Radecki, Effects of imidocarb dipropionate in cats with chronic haemobartonellosis, Vet. Ther. 2 (2002) 144e149. S. Tasker, C.R. Helps, M.J. Day, D.A. Harbour, T.J. Gruffydd-Jones, M. Lappin, Use of a Taqman PCR to determine the response of Mycoplasma haemofelis infection to antibiotic treatment, J. Microbiol. Methods 56 (2003) 63e71. K.L. Dowers, C. Olver, S.V. Radecki, M.R. Lappin, Use of enrofloxacin for treatment of large-form Haemobartonella felis in experimentally infected cats, J. Am. Vet. Med. Assoc. 221 (2002) 250e253. S. Tasker, C.R. Helps, M.J. Day, T.J. Gruffydd-Jones, D.A. Harbour, Use of real-time PCR to detect and quantify Mycoplasma haemofelis and ‘Candidatus Mycoplasma haemominutum’ DNA, J. Clin. Microbiol. 41 (2003) 439e441. J.J. Callanan, H. Thompson, S.R. Toth, B. O’Neil, C.E. Lawrence, B. Willett, O. Jarrett, Clinical and pathological findings in feline immunodeficiency virus experimental infection, Vet. Immunol. Immunopathol. 35 (1992) 3e13. A.J. German, M.J. Cannon, C. Dye, M.J. Booth, G.R. Pearson, C.A. Reay, T.J. Gruffydd-Jones, Oesophageal strictures in cats associated with doxycycline therapy, J. Feline Med. Surg. 7 (2005) 33e41. Y.L. McGrotty, C.M. Knottenbelt, Oesophageal stricture in a cat due to oral administration of tetracyclines, J. Small Anim. Pract. 43 (2002) 221e223. L.D. Melendez, D.C. Twedt, M. Wright, Suspected doxycycline-induced oesophagitis with oesophageal stricture formation in three cats, Feline Pract. 28 (2000) 10e12.

S. Tasker et al. / Microbes and Infection 8 (2006) 653e661 [24] K.N. Gelatt, A. van der Woerdt, K.L. Ketring, S.E. Andrew, D.E. Brooks, D.J. Biros, H.M. Denis, T.J. Cutler, Enrofloxacinassociated retinal degeneration in cats, Vet. Ophthalmol. 4 (2001) 99e106. [25] J.C. Flint, M.H. Roepke, R. Jensen, Feline infectious anaemia. II. Experimental cases, Am. J. Vet. Res. 20 (1959) 33e40.

661

[26] J.E. Woods, J.R. Hawley, M.R. Lappin, Attempted transmission of Haemobartonella felis by Ctenocephalides felis, J. Vet. Int. Med. 17 (2003) 426. [27] S.E. Shaw, M.J. Kenny, S. Tasker, R.J. Birtles, Pathogen carriage by the cat flea Ctenocephalides felis (Bouche´) in the United Kingdom, Vet. Microbiol. 102 (2004) 183e188.