Development and validation of 2 probe-hybridization quantitative PCR assays for rapid detection of a pathogenic Coxiella species in captive psittacines

Coxiella spp. (phylum Proteobacteria, class Gammaproteobacteria, order Legionellales, family Coxiellaceae) are gram-negative, obligate intracellular bacteria that infect the phagolysosomes of macrophages. ⁸ This genus is most known for the species Coxiella burnetii, the causative agent of Q fever. This bacterium causes significant disease in humans and animals worldwide and is commonly spread by direct contact or aerosol inhalation of infected body fluids associated with parturition in ruminants. ¹⁹ In humans, coxiellosis can appear with a variety of nonspecific signs; however, chronic infection may result in endocarditis (https://www.cdc.gov/qfever/symptoms/index.html). Infection in ruminants is usually subclinical but can appear as anorexia and late-term abortion. ¹⁹

Diagnosis of coxiellosis relies on serology and PCR detection of the agent. ⁸ Real-time PCR is superior for early diagnosis of Q fever in humans. ¹¹ Treatment yields variable results and includes 18−24-mo courses of doxycycline for chronic disease (https://www.cdc.gov/qfever/symptoms/index.html). ⁷ Although ruminants may be the most recognized reservoirs of C. burnetii, several other mammals and birds are possible hosts.^5,16 The zoonotic and bioterrorism potential of C. burnetii warrants its status as a reportable disease in the United States (https://www.niaid.nih.gov/research/emerging-infectious-diseases-pathogens).

Until recently, C. burnetii was the sole species within the genus Coxiella. However, the identification of Coxiella cheraxi in crawfish ¹⁵ and other Coxiella-like endosymbionts in > 40 species of ticks indicates that greater diversity exists. ¹⁹ A study investigating the evolutionary origin of the genus Coxiella highlights 5 genes (rpoB, 16s rRNA, 23s rRNA, groEL, dnaK) found to be conserved within samples from ticks positive for Coxiella infection, ³ and reports that all C. burnetii strains could be traced back to a Coxiella-like progenitor hosted by ticks.

Although birds do not appear to develop clinical disease from exposure to C. burnetii, ¹³ other Coxiella spp. appear to be more pathogenic. A previously undescribed and currently unnamed species of Coxiella (AvCox) has been implicated in the death of individuals of several species of psittacines and one toucan to date.^6,14,17,18 These birds had nonspecific clinical signs including lethargy, weakness, and neurologic signs. Hematologic abnormalities, if present, indicated chronic infection. Imaging indicated various degrees of splenomegaly, hepatomegaly, or cardiac changes. Postmortem examinations revealed various levels of emaciation, splenomegaly, and hepatomegaly. Histologic findings included Coxiella-like bacteria within macrophages of the spleen, liver, bone marrow, kidneys, and adrenals. Other lesions included granulomatous encephalitis and myocarditis. This bacterium was confirmed using 16S bacterial PCR and electron microscopy. ¹⁴ In all but one case, the causative agent was identified postmortem.

Avian coxiellosis may be an emerging and/or underdiagnosed disease of captive birds. Because clinical signs are nonspecific and there is currently no validated antemortem test, diagnosis of this disease has often been made postmortem. Rapid and specific testing, such as quantitative PCR (qPCR), is essential for the early diagnosis and treatment of avian coxiellosis.

We collected samples from 4 birds that had previously been confirmed positive for AvCox via bacterial 16S PCR and sequencing. Antemortem samples included conjunctival, cloacal, and choanal swabs as well as whole blood from a blue-and-gold macaw (Ara ararauna). Postmortem samples included spleen from the blue-and-gold macaw (Fig. 1), ⁶ liver and spleen from a scarlet-chested parrot (Neophema splendida), and a combined liver–spleen sample from a Timneh African grey parrot (Psittacus timneh) (Fig. 2). Additionally, whole blood, splenic tissue, and conjunctival, cloacal, and choanal swabs were collected from a blue-and-gold macaw of unknown infection status that had been housed with the positive macaw. Liver from a yellow-naped Amazon parrot (Amazona auropalliata) that was AvCox negative on 16S rRNA PCR was used as a negative control.

OPEN IN VIEWER

Figure 1.

Blue-and-gold macaw, case 1. The spleen (Spl) is markedly enlarged (5.2 × 3.5 × 3.5 cm) and microscopically contained infarcts and hemorrhage. The liver (Liv) is mottled pale and slightly enlarged. No gross lesions are noted in the heart (Hrt), lung (Lung), proventriculus (Pvt), ventriculus (Ven), or pancreas (Pan). Bar = 2 cm. Figure 2. African grey parrot, case 5. Bile-tinged liver with moderate diffuse enlargement and rounded edges. Figure 3. African grey parrot, case 5. The cytoplasm of hepatic macrophages is markedly expanded by smudgy blue granular content (demarcated by arrowheads). H&E. Bar = 10 μm. Figure 4. African grey parrot, case 5. Hepatic macrophages are distended by phagolysosomes containing purple-stained bacteria on a densely eosinophilic background (demarcated by arrowheads). Gimenez. Bar = 10 μm.

DNA was extracted from all samples (DNeasy blood and tissue kit; Qiagen, Germantown, MD), and DNA concentration was measured (NanoDrop 1000 spectrophotometer; Thermo Fisher, Gaithersburg, MD). All extracts were aliquoted and stored at −80°C. To confirm the status of the samples, nested PCR amplification of the bacterial 16S rRNA gene was performed using consensus primers as described previously. ¹² The PCR product was resolved in 1% agarose gel, the band was excised, and DNA was extracted (QIAquick gel extraction kit; Qiagen). Gene products were labeled (BigDye terminator kit; Applied Biosystems, Foster City, CA) and sequenced by Sanger methodology.

Three of the 5 genes described as conserved within Coxiella-positive samples ³ were used to obtain the template sequence for qPCR design. Genes dnaK (70-kDa heat shock protein), groEL (chaperone protein), and rpoB (RNA polymerase beta subunit) were amplified using pan-Coxiella primers as described previously.^3,4 Products were electrophoresed, extracted, and sequenced as above.

The obtained AvCox dnaK, groEL, and rpoB sequences were used to design sets of primers and probes for each gene, using MAFFT v.7 (https://mafft.cbrc.jp/alignment/software/) and Primer3Plus (http://www.bioinformatics.nl/cgi-bin/primer3plus/primer3plus.cgi). The first set included forward primer CoxGroELF (5’-TGCAATTTGATCGTGGCTAT-3’), reverse primer CoxGroELR (5’-AACGGATCTTCGAGTTCTGC-3’), and probe CoxGroELprobe (FAM-TGTCACCATATTTTATCAACAACCAGCAAAA-BHQ1). The second included forward primer CoxdnaKF (5’-ACAACGCCTGAAAGAAGCAT-3’), reverse primer CoxdnaKR (5’-TGGCAAGTTAACATCCGTTT-3’), and probe CoxdnaKprobe (TX Red-AAGCGAAAATTGAGCTTTCTTCTAGCCAA-BHQ2). The third included forward primer CoxrpoBF (5’-TCGCAAAGGTGTCCCTATATC-3’), reverse primer CoxrpoBR (5’-TTGACCAGAGGTTGGGAGTC-3’), and probe CoxrpoBprobe (Cyanine3-TCACCTGTATTTGATGGTGCACTGAA-BHQ2).

All qPCR reactions took place on a 96-well reaction plate (MicroAmp Fast optical reaction plates; Applied Biosystems). Each reaction well consisted of a 20-μL solution containing 1 μL each of forward primer, reverse primer, and probe; 7 μL of sample; and 10 μL of TaqMan master mix (TaqMan Fast universal PCR master mix 2×; Applied Biosystems). Each reaction was run in duplicate. A eukaryotic 18S rRNA endogenous control kit (VIC/MGB probe; Applied Biosystems) was used to confirm the presence of amplifiable DNA in each sample. Nuclease-free water was used in 6 wells to serve as non-template controls.

The standard curve was created using 10-fold serial dilutions of the AvCox amplicon. Control standards were run using 10¹–10⁶ copies in duplicate on each plate. The reactions were amplified (7500 Fast real-time PCR system; Applied Biosystems). Cycling conditions included initial denaturation at 95°C for 20 s, 40 cycles of 95°C for 3 s, then 60°C for 30 s. The function of the qPCR probes and primers was assessed by evaluating R² values and slopes of the standard curves, calculated using the software included with the ABI 7500 fast equipment. The correlation between the groEL cycle threshold (Ct) values and the dnaK Ct values was assessed using a Wilcoxon rank-sum statistical analysis (Prism 8; GraphPad Software, San Diego, CA).

Postmortem and microscopic examination of 3 patients were conducted at the University of Florida (Gainesville, FL; A. ararauna), Alabama Department of Agriculture and Industries (Montgomery, AL; P. timneh), and Zoo/Exotic Pathology Service (Carmichael, CA; N. splendida), and involved 5 pathologists using various methods including routine hematoxylin and eosin staining, special histologic stains, including Gimenez stain, and transmission electron microscopy (Figs. 3-6). Techniques and staining protocols varied among institutions, and autopsy findings and final diagnoses of avian coxiellosis were obtained retrospectively from written reports. Variables associated with diagnostic methodology were not controlled.

OPEN IN VIEWER

Figure 5.

Liver of case 6, scarlet-chested parrot. Phagolysosomes in macrophages contain clusters of bacteria 0.4–0.7 μm diameter. Bar = 1.0 µm. Figure 6. Bacteria have a double-membrane structure (arrowheads) with inner cytoplasmic membrane and outer membrane or cell wall separated by periplasmic space. Transmission electron microscopy. Bar = 0.5 µm.

The AvCox-positive birds had a variety of nonspecific clinical signs (Table 1). Laboratory testing revealed leukocytosis, a left shift with toxic change, elevated creatine kinase (CK) activity, and splenomegaly and hepatomegaly (Table 1). All birds in our study had histiocytic hepatitis. Splenomegaly was observed in 2 birds (Table 2). For birds in which further testing was not performed or additional information was unavailable, results are listed as “unknown.” Because Chlamydia psittaci infection is also consistent with these findings, all samples were tested for C. psittaci by PCR as described previously ² ; samples were uniformly negative.

OPEN IN VIEWER

Table 1.

Clinical findings in 5 Coxiella-positive psittacines.

Species	Clinical signs	Laboratory findings	Imaging findings
Ara ararauna (cases 1, 2)	Lethargy, poor body condition, decreased range of motion in both humeroulnar joints.	Left shift, 10 toxic change.	Splenomegaly, hepatomegaly.
Psittacus timneh (case 5)	Lethargy and weakness of 2.5-mo duration, weight loss.	Unknown.	Unknown.
Neophema splendida (cases 6, 7)	Lethargy, poor body condition, scanty droppings.	Mild leukocytosis, elevated CK. 9	Unknown.

CK = creatine kinase. Cases 3 and 4 were negative for AvCox and were excluded from this table (case 3 was qPCR negative, and case 4 was 16S PCR and qPCR negative).

OPEN IN VIEWER

Table 2.

Postmortem findings in 5 Coxiella-positive psittacines.

Species	Gross findings	Histologic findings	Additional testing
Ara ararauna (cases 1, 2)	Splenomegaly, hepatomegaly.	Histiocytic splenitis; intrahistiocytic PAS-positive material in macrophages of spleen, ventriculus, liver, intestine, and pancreas.	Not conducted.
Psittacus timneh (case 5)	Hepatomegaly, bile-tinged liver, coelomitis.	Hepatitis, with intracellular Gimenez-positive intracellular bacteria, pneumonia, coelomitis, airsacculitis. Spleen was not examined microscopically but was pale and normal sized grossly.	Electron microscopy: clusters of variably sized bacteria 0.4–0.7 µm diameter.
Neophema splendida (cases 6, 7)	Splenomegaly, hepatomegaly, cerebral hemorrhage.	Lymphoplasmacytic histiocytic splenitis and hepatitis; granulocytic extramedullary hematopoiesis.	Electron microscopy: clusters of small membrane-bound bacteria.

After primer sequences were edited out, the dnaK PCR resulted in a product of 466 bp, the groEL PCR resulted in a product of 575 bp, and the rpoB PCR resulted in a product of 491 bp. BLASTn searches for each sequence found that they were most homologous to Coxiella endosymbionts of ticks (top hit for dnaK, 87% homology to GenBank accession KP985416; top hit for groEL, 87% homology to GenBank accession KP985504; top hit for rpoB, 87% homology to GenBank accession KP985331). Sequences were submitted to GenBank under accessions MK156101–MK156103.

The 3 assays were initially run together on the same plate as a multiplex assay; however, efficiency was unacceptable. Each target was then run as a single assay. The standard curve of the groEL assay had a slope of −3.49, indicating 93.4% efficiency, and R² of 0.99. The standard curve of the dnaK assay had a slope of −2.99, indicating 116% efficiency, and R² of 0.99. The rpoB assay performed poorly and was not used further. There was 100% agreement of the groEL and dnaK qPCR assays with the results of the consensus PCR; this consisted of 4 positive and 1 negative sample. There was 100% agreement between positive and negative status for all samples with the groEL and dnaK assays. A Wilcoxon rank-sum test found a strong correlation between groEL Ct values and dnaK Ct values (p < 0.001; Table 3). There was 100% agreement between our 2 qPCR assays and the previously available 16S rRNA gene amplification and sequencing assay, consistent with 100% sensitivity and specificity. Larger sample sizes may have found inconsistencies. The correlation between the quantities detected in the 2 qPCR assays was very strong, as determined by the Wilcoxon signed-rank test.

OPEN IN VIEWER

Table 3.

Quantitative PCR results for groEL and dnaK assays.

Case	Species	Tissue	Consensus PCR	groEL Ct	dnaK Ct
1	Ara ararauna	Spleen	Positive	27.8 (1.8)	29.6 (3)
2	Ara ararauna	Spleen	Positive	27.7 (0.18)	29.4 (0.1)
		Choanal swab	Not tested	33.8 (0.4)	34.1 (0.1)
		Conjunctival swab	Not tested	33.7 (2.5)	31.2 (6.3)
		Cloacal swab	Not tested	24.9 (0.3)	26.0 (0.5)
		Whole blood	Not tested	38.0 (UND)	37.1 (0.7)
3	Ara ararauna	Choanal swab	Not tested	Negative	Negative
		Conjunctival swab	Not tested	Negative	Negative
		Cloacal swab	Not tested	Negative	Negative
		Whole blood	Not tested	Negative	Negative
4	Amazona auropalliata	Liver	Negative	Negative	Negative
5	Psittacus timneh	Liver	Positive	22.6 (0.2)	25.5 (0.4)
6	Neophema splendida	Combined liver–spleen	Positive	25.3 (0)	24.9 (0.2)
7	Neophema splendida	Combined liver–spleen	Positive	25.3 (0)	25.0 (0)

Ct = cycle threshold; UND = undetermined. Numbers in parentheses are standard deviations.

Splenic and hepatic samples consistently yielded high counts of genetic material for this Coxiella species when using our qPCR assays. Because these samples were collected either postmortem or via ultrasound-guided, fine-needle aspirate, a less invasive sample is desirable for more routine use of these assays in clinical practice. Cloacal, conjunctival, and choanal swabs, as well as whole blood, had detectable amounts of DNA from one positive bird, with the cloacal swab yielding the highest amount and blood yielding the lowest amount. Larger scale studies with more birds are needed to determine whether these differences are significant.

With the identification of this novel Coxiella species that has caused fatal disease in multiple cases, AvCox warrants consideration as an emerging or previously under-recognized disease of captive birds. Its close relative, C. burnetii, may serve as a model for predicted behavior. It is likely that AvCox has the potential to be infectious between individuals, either through direct contact or aerosol inhalation of infected body fluids. Additionally, it has been theorized that C. burnetii likely evolved from a tick endosymbiont ³ ; multiple Coxiella-like endosymbionts have been found in > 40 species of ticks, including ticks of seabird and psittacine colonies.^1,3,4 It is unknown how the birds in our cases contracted this bacterium. All birds in our study were housed outdoors, making exposure to arthropod vectors likely. Direct contact with body fluids from infected individuals would seem like a likely source of infection; however, 2 of the blue-and-gold macaws in our study are surprising. Cases 2 and 3 were caged together, and although one had a fatal AvCox infection, the other had no evidence of infection found on autopsy or histopathology and was qPCR negative.

Clinical signs of AvCox infection in our cases included lethargy, poor body condition, decreased range of motion in humeroulnar joints, and scanty droppings. Blood work revealed leukocytosis, left shift with toxic change, and elevated CK activity. Diagnostic imaging revealed splenomegaly and hepatomegaly. Many of these findings are consistent with previously reported cases of AvCox infection.^6,14,17,18 Of these, splenomegaly was the most consistent finding across species and was observed in 2 of 3 birds autopsied in our study. A left shift with toxic change was consistent in the infected blue-and-gold macaws (cases 1, 2). Elevated CK appears to be a new finding and was only found in the scarlet-chested parakeets (cases 6, 7); however, the relevance of this finding in the context of the AvCox disease process is unknown. The African grey parrot in our report had marked hepatic macrophage intracellular accumulation of AvCox bacteria easily definable with Giemsa staining that could represent an infection of greater chronicity or may demonstrate species-specific variation in AvCox pathogenesis.

In addition to the clinical importance of AvCox for avian patients, the potential zoonotic risk merits investigation. Although transmission of avian coxiellosis to people has not been documented, the similarities of this Coxiella species to C. burnetii justify proper biosecurity measures when dealing with suspected cases.