Anemia Associated With ’Candidatus Mycoplasma haemominutum’ in a Feline Leukemia Virus-Negative Cat With Lymphoma

Case Report

A 16.5-year-old, 3.75-kg, castrated male domestic shorthair cat presented to the University of Illinois Veterinary Teaching Hospital (UI-VTH) for evaluation of anemia of unknown duration and for ptyalism. The indoor/outdoor cat was vaccinated regularly against feline leukemia virus (FeLV). In the past, the cat tested negative with a combined commercial enzyme-linked immunosorbent assay (ELISA) for detection of FeLV and feline immunodeficiency virus (FIV). Abnormalities detected on physical examination included a dull hair coat accompanied by flea infestation, prior enucleation of the right eye (6 years previously, from trauma), pale-pink mucous membranes, periodontal disease, a mass under the tongue, and submandibular lymphadenopathy. The dental disease and sublingual lesion were the suspected causes of ptyalism and lymphadenopathy.

Initial laboratory tests included a complete blood cell count (CBC), serum biochemical profile, urinalysis and urine culture, and total serum thyroxine (T4). The only abnormalities found were a mild, normochromic-normocytic anemia (red blood cell count [RBC] 6.22 × 10⁶/μL; reference range, 5.00 to 10.00 × 10⁶/μL and hematocrit [HCT] 24.4%; reference range, 30% to 45%). Reticulocyte counts were low, with 7% punctate reticulocytes (435 × 10³/μL) and 0.3% aggregate reticulocytes (18 × 10³/μL), indicating a nonregenerative anemia. Thoracic radiography and abdominal ultrasonography were also performed. The thoracic radiographs (i.e., ventrodorsal, right and left lateral) did not reveal any abnormalities. The abdominal ultrasound showed a hyperechoic mass measuring 14.9 × 13.2 mm, arising from the right side of the liver, as well as a spleen of heterogeneous mixed echogenicity. Hyperechoic areas in the left kidney consistent with age-related chronic changes, and echogenic material in the urinary bladder consistent with crystals were also observed. The cat was discharged with a prescription of fipronil^a to control the existing flea problem, and the owner was instructed to return the cat as soon as possible for biopsies of the oral lesion and submandibular lymph nodes.

Twelve days later, excisional biopsy of the sublingual lesion and incisional biopsy of the left submandibular lymph node were obtained. Histopathology showed a severe, chronic-active ulcerative plasmacytic stomatitis with epithelial hyperplasia in the mouth and a diffuse, large-cell lymphoma within the lymph node. Immunohistochemistry revealed the lymphoma to be of B-cell lineage (CD79a positive, CD3 negative).

On the third evaluation 8 weeks after the initial presentation, a second abdominal ultrasound failed to reveal any significant new findings. Cytology of a fine-needle aspirate of the spleen revealed a heterogenous cell population consisting of 70% small, well-differentiated lymphocytes and 30% lymphoblasts, which was not conclusive for lymphoma. However, cytological evaluation of a mildly enlarged popliteal node showed a monomorphic population of large lymphoblasts with multiple nucleoli, confirming the presence of lymphoma. The previously described liver mass could not be observed on the second ultrasound. A combined ELISA for FIV antibody and FeLV antigen^b was negative. A CBC showed persistent mild normochromic-normocytic anemia with an RBC count of 6.68 × 10⁶/μL (HCT, 27.6%), and there were no significant changes in the serum biochemical profile.

A diagnosis of FeLV-negative, multicentric, nodal, diffuse lymphoblastic B-cell lymphoma with potential splenic involvement was made, and therapy was instituted. A multi-agent protocol was initiated, incorporating an initial subcutaneous dose of L-asparaginase^c (400 U/kg), followed by weekly to biweekly doses of intravenous (IV) vincristine^d (0.4 mg/m², initial dose) interspersed with IV doses of actinomycin-D^e (0.54 to 0.58 mg/m²) on weeks 10, 14, and 18. The cat also received low-dose prednisone^f (20 mg/m² per os [PO] once daily initially and gradually decreased to 10 mg/m² every other day by week 4), as well as cyclophosphamide^g (25 mg PO every 5 to every 7 days). Cyclophosphamide was changed to chlorambucil^h following the initial 8 weeks of therapy (2 mg PO q 5 to 7 days). After the 24th week of therapy, vincristine administration was decreased to every 3 weeks. The cat was in complete clinical remission by week 3 of treatment, based on no palpable lymphadenopathy or splenomegaly. Because a CBC was performed before each administration of IV chemotherapeutic agents, evolution of the anemia was closely monitored. The RBC count constantly remained between 5.94 and 7.58 × 10⁶/μL (HCT, 25% to 30%) after initiation of therapy, with values of 6.40 to 7.00 × 10⁶/μL most commonly obtained (HCT, 27% to 29%). Flea infestation was a continual problem.

On week 22 of therapy, the cat was presented with a decreased appetite. At this time, the RBC count was 4.50 × 10⁶/μL, down from 7.11 × 10⁶/μL 2 weeks prior (HCT, 19.1%; down from 28.2%). Examination of a blood smear stained with Giemsa-Wright revealed one or two small, cocci-shaped parasites on approximately 25% of the erythrocytes [Figure 1]. An in-house polymerase chain reaction (PCR) assay detecting specific 16S rRNA sequences was performed to identify the exact nature of the hemotrophic parasites. Treatment for presumptive hemobartonellosis was started with doxycyclineⁱ (6.9 mg/kg PO q 12 hours for 21 days), and cyproheptadine^j (2 mg PO q 12 hours) was initiated as an appetite stimulant. The results of the PCR assay showed a strong signal for ‘Candidatus Mycoplasma haemominutum’ (i.e., Haemobartonella felis, California species/small form) and were negative for Mycoplasma haemofelis (i.e., H. felis, Ohio species/large form).

Seven days later, the cat was presented with lethargy, anorexia, decreased water intake, and difficulty defecating. On physical examination, the cat’s mucous membranes were pale. Clinical dehydration was estimated to be 8%, and a small amount of firm feces was palpated in the rectum. The cat was still in complete remission for its lymphoma, with no palpable lymphadenopathy or organomegaly. Despite therapy with doxycycline and the presence of dehydration, the RBC count had fallen to 3.83 × 10⁶/μL (HCT, 16.7%). A serum biochemical profile and urinalysis were within reference ranges. Lower numbers of the hemotrophic parasite were observed on blood smears, and a second PCR was still positive (weaker signal) for ‘Candidatus Mycoplasma haemominutum’ [Figure 2], probably from owner noncompliance with the administration of the doxycycline. The cat was hospitalized and placed on IV lactated Ringer’s solution fluids to correct dehydration. Oral doxycycline therapy and IV ranitidine^k (1 mg/kg q 12 hours) were also administered. Within 36 hours, the cat began eating and appeared bright and alert. No blood products were administered. The cat was discharged with instructions for the owner to discontinue chlorambucil until further notice, as this bifunctional alkylator is an immunosuppressive agent. The dose of prednisone was increased to 20 mg/m² once daily.

The following week, the cat was still alert and well hydrated with pale-pink mucous membranes. A CBC revealed that the RBC count had increased to 5.96 × 10⁶/μL (HCT, 24.6%). A third PCR was negative, and no organisms were found on peripheral blood smears. The cat was treated with vincristine on that visit, and the doxycycline and prednisone were continued at the same dosages. During this visit, blood was also collected from two healthy housemates and submitted for CBC and PCR analysis. The RBC counts (>8.50 × 10⁶/μL; HCT, >40%) were within the expected reference range for both cats, and no parasites were visualized on the peripheral blood smears. One of the housemates was positive by PCR for ‘Candidatus Mycoplasma haemominutum,’ while the other one [Figure 2] was negative. The positive housemate had been hospitalized and treated, 9 months prior, for an episode of seizures suspected to be caused by a Toxoplasma spp. infection that responded to clindamycin therapy. Four live fleas were also captured and submitted for PCR analysis [three fleas from the PCR-negative cat and one flea from the PCR-positive cat]. They were homogenized, and a PCR was performed. The test revealed a positive, weak signal for ‘Candidatus Mycoplasma haemominutum’ [Figure 2]. The PCR products were gel purified and sequenced in the sense and antisense directions, by use of a dideoxy terminator method. These sequences were identical to one another and to the ‘Candidatus Mycoplasma haemominutum’ previously deposited in GenBank.¹ All cats and fleas were negative for Mycoplasma haemofelis on PCR [Figure 2].

At the following six visits to the UI-VTH, the cat’s RBC counts remained between 4.90 and 7.20 × 10⁶/μL (HCT, 21% to 27%). The cat remained in complete clinical remission on vincristine every 3 weeks and prednisone every other day. At week 44, the cat was acutely presented with lethargy, mild dehydration, cyanosis, and neurological signs that progressed rapidly following admission. No peripheral lymphadenopathy or organomegaly was detected. On a CBC the RBC count was 7.39 × 10⁶/μL (HCT, 27%), and no parasites were detected on blood smears. A PCR assay was weakly positive, however. A complete serum biochemical profile and urinalysis failed to identify significant abnormalities. Because of the dramatic worsening of the cat, the owner elected euthanasia but declined a complete necropsy.

Discussion

Feline infectious anemia was first described in South Africa 60 years ago.² The organism responsible was initially named Eperythrozoon felis, but the name was changed to Haemobartonella felis in 1955.² The organism was originally classified in the family Anaplasmataceae, order Rickettsials, on the basis of its gram-staining characteristics, small size and obligate parasitic nature, susceptibility to tetracyclines, and inability to cultivate the organisms in vitro.²³ Transmission of the parasite may occur via bloodsucking arthropods, such as fleas and ticks, although confirmatory studies of horizontal vector transmission are lacking.²³ One of the major difficulties in studying the organisms is that they cannot be successfully cultivated outside the host.²³ Until the recent development of PCR and real-time PCR (RT PCR)-based assays to detect specific fragments of the parasites’ nucleic acid sequence, the only test available for diagnosis was manual evaluation of a blood smear.^1–10 The latter test is fraught with problems, such as poor sensitivity and specificity, making it unreliable for establishing the presence of an infection.

Two different “strains” of causative organisms have been identified to date, using the PCR or RT PCR techniques.^124–10 Phylogenic studies suggest that these parasites are not just different strains, but truly distinct species.^127–13 The diseases caused by these two hemotrophic parasites in the cat are also distinctly different.^128–1013 In addition, sequencing of their 16S rRNA genes indicates a closer relationship to the family Mycoplasmataceae, which has necessitated reclassification and renaming of these organisms.^19–12 The large form of organism (H. felis, Ohio species) is now known as Mycoplasma haemofelis, whereas the smaller form (H. felis, California species), an incompletely characterized species, has been given the temporary designation ‘Candidatus Mycoplasma haemominutum.’^19–12 The causal relationship between infection with Mycoplasma haemofelis and the acute form of feline infectious anemia was recently confirmed.¹⁵

Coinfection with H. haemofelis and retroviruses is common, and an association between FeLV infection and more severe clinical disease from Haemobartonella has been documented.^31316–18 The cat in this report was tested twice for retroviruses with an ELISA and was negative on both occasions.

The clinical signs caused by Mycoplasma haemofelis include anorexia, lethargy, weight loss, pale mucous membranes, icterus, and fever.²³ In severe cases of overwhelming parasitemia, acute hemolytic anemia may develop and result in death.²³ Healthy cats that were experimentally infected solely with ‘Candidatus Mycoplasma haemominutum’ developed only minimal signs of disease and typically had HCT values >30%.⁸⁹¹³¹⁴ These observed differences in virulence may be merely dose-dependent effects. The possible sequelae of chronic infections with these hemotrophic parasites have not been defined in cats, and it cannot be assumed that these infections are inconsequential. Recently, a study described more potential for significant anemia in cats coinfected with ‘Candidatus Mycoplasma haemominutum’ and retroviruses.¹³ The authors speculated that the mycoplasmal agent may have been a factor in the higher-than-expected incidence of myeloproliferative disorders in the cats of that study.¹³

The traditional therapy for feline infectious anemia includes the administration of antibiotics and supportive therapy with blood products in severely anemic cats.²³ The administration of immunosuppressive doses of corticosteroids has also been recommended, especially in cases when an immune-mediated component is suspected, or in cases with an acute onset, autoagglutination, or a positive Coombs’ test.²³ The cat reported here had a more insidious onset of signs, different from infection with the large form (Mycoplasma haemofelis). However, the decision was made to discontinue chlorambucil, an immunosuppressant that causes apoptosis of T and B lymphocytes, and the dosage of prednisone was increased in an effort to decrease erythrophagocytosis by the reticuloendothelial system. Treatment with tetracyclines, such as doxycycline, has been effective in abolishing patent infections by Mycoplasma haemofelis and ‘Candidatus Mycoplasma haemominutum’ in cats, although sustained clearance of organisms, confirmed by PCR, has only been rarely reported.²³⁵⁸ A recent trial with azithromycin^l showed that this newer macrolide antibiotic was of no clinical benefit for the treatment of either form of the parasite.¹⁴ In contrast, efficacy of enrofloxacin^m against Mycoplasma haemofelis was demonstrated when used at a dose of 10 mg/kg q 24 hours for 14 days.¹⁹²⁰ The effectiveness of enrofloxacin for treatment of infections with the small form of the parasite has not yet been determined. Additional studies have shown that imidocarb dipropionateⁿ failed to consistently clear ‘Candidatus Mycoplasma haemominutum’ from chronically infected cats.²¹ While imidocarb dipropionate was shown to be relatively safe in cats, it is not known whether treatment permanently clears organisms during acute infections, thereby preventing the development of a chronic infection. The use of imidocarb for acute infections warrants further investigation, with larger numbers of experimentally infected cats.

Although healthy cats infected with ‘Candidatus Mycoplasma haemominutum’ may fail to develop a clinically significant anemia, antibiotic treatment may still be indicated. It may be necessary to treat affected cats to help prevent infection of other cats in the household. The possibility that recrudescence of parasitemia may lead to acute disease, or that other diseases may arise with chronic infections, suggests that treatment of any infected cat should be considered. On the other hand, it could be argued that therapy does not clear the infection, and, therefore, it is not necessary to treat a cat that is positive but shows no clinical sign of the disease (e.g., housemate cat with a Hct of 42%). One risk of not treating the positive housemate was that the cat reported here with lymphoma could have been reinfected from the infected housemate. In this case, a decision was made to treat only the clinically ill cat. When retested on the day of euthanasia, the cat reported here was weakly positive on PCR assay but had no further decrease in his HCT. The two clinically normal housemates were retested with a PCR assay and a CBC 9.5 months after their initial tests, and the results were identical. Both of the cats still had live fleas. Perhaps the healthy negative cat was able to clear the infection. However, it is also possible that the healthy cat had extremely low levels of parasitemia or that the parasites were sequestered in certain tissues or organs, making them undetectable by PCR on peripheral blood.

Conclusion

This case report illustrates that a clinically significant, life-threatening anemia may develop in an older cat infected with ‘Candidatus Mycoplasma haemominutum’ if there is concurrent disease (e.g., lymphoma), stress (e.g., frequent hospital visits and fleas), and/or immunosuppression (e.g., prednisone and alkylator therapy). One of two other cats sharing the house with the cat reported here was infected but did not show clinical signs and did not become anemic. This latter finding was consistent with previous reports that suggest infection with ‘Candidatus Mycoplasma haemominutum’ is unlikely to cause overt anemia in otherwise healthy cats.

Acknowledgment

The authors thank Therese E. Eggett for technical assistance in performing the PCR assays.