CHANGES IN TISSUE TROPISM OF MYCOPLASMA GALLISEPTICUM FOLLOWING HOST JUMP

Mycoplasma gallisepticum is a pathogen of worldwide economic importance, particularly for the poultry industry. In chickens, the bacterium is recovered, especially from the respiratory tract (Levisohn and Kleven 2000), where it typically causes chronic respiratory disease, including inflammation of the air sacs, lungs, and trachea (Pflaum et al. 2020). In turkeys, it frequently causes infraorbital sinusitis (Ferguson 2013; Raviv and Ley 2013). Extensive variability in tissue tropism has been reported (Levisohn and Kleven 2000), including isolation of encephalitic forms in turkeys (Thomas et al. 1966; Chin et al. 1991), strains with a proclivity for the cloacal tissue (Macowan et al. 1983), strains causing unilateral enlargement of the eyeball (Power and Jordan 1976), and rare cases of keratoconjunctivitis in layer chickens (Nunoya et al. 1995). Around 1993–94, an epidemic of mycoplasmal conjunctivitis in wild birds, especially House Finches (Haemorhous mexicanus), emerged in the Baltimore, Maryland, area (Doster 1994), caused by a novel strain of M. gallisepticum (Ley et al. 1997; Hochachka et al. 2013). The epidemic spread rapidly across eastern North America (Fischer et al. 1997; Dhondt et al. 1998) and later also spread westward causing disease, especially in House Finches (Ley et al. 2006). In finches, clinical signs and gross lesions ranged from mild to severe unilateral or bilateral conjunctival swelling with serous to mucopurulent drainage and nasal exudate. Microscopic lesions consisted of chronic lymphoplasmacytic conjunctivitis, rhinitis, and sinusitis (Ley et al. 1996).

Given that M. gallisepticum disease in poultry is primarily associated with the upper respiratory tract, the traditional location for sampling in most poultry studies is the trachea or choanal cleft (e.g., Feberwee, Mekkes, de Wit et al. 2005; Feberwee, Mekkes, Klinkenberg 2005; García et al. 2005; Feberwee et al. 2006; Gharaibeh and Hailat 2011; Pflaum et al. 2017). Similarly, older studies in wild birds sampled and detected M. gallisepticum in the trachea, for example, in House Sparrows, Passer domesticus (Jain et al. 1971) and Tree Sparrows, Passer montanus (Shimizu et al. 1979). After the emergence of M. gallisepticum in wild birds in North America, most studies continued to use choanal swabs (Farmer et al. 2002, 2005; Ganapathy et al. 2007) or a pooled sample of choanal and conjunctival swabs (e.g., Luttrell et al. 1996, 1998, 2001; Fischer et al. 1997; Staley et al. 2018; Tardy et al. 2019; Bonneaud et al. 2020). Few studies have sampled and analyzed conjunctival and choanal samples separately (e.g., Michiels et al. 2016). In our finch studies, we historically sampled the conjunctiva rather than the choana (Hartup and Kollias 1999; Hartup et al. 2000, 2001; Altizer et al. 2004; Dhondt et al. 2005, 2007, 2008, 2013, 2015, 2017; Sydenstricker et al. 2006; Hawley et al. 2010; Fleming-Davies et al. 2018). In the one study in which we took both choanal and conjunctival swabs from House Finches, M. gallisepticum was detected separately at both locations up to 20 wk after inoculation (Kollias et al. 2004) by PCR following the Lauerman (1998) protocol.

The lack of attention to a possible effect of the sampling location on the detection and bacterial load of M. gallisepticum is illustrated in a recent meta-analysis identifying wild bird species in which M. gallisepticum has been detected (Sawicka et al. 2020). The results compiled in that study were grouped using multiple criteria that did not, however, include the sampling location in the host.

It is evident that the clinical signs following a M. gallisepticum infection differ markedly between poultry and House Finches; the M. gallisepticum strain that emerged in the early 1990s in House Finches clearly forms a separate phylogenetic clade (HOFI_MG) from the POULTRY_MG clade causing disease in poultry (Ley et al. 1997; Hochachka et al. 2013); and gene expression in the same tissues using the same M. gallisepticum strain differed between poultry and House Finches (Pflaum et al. 2020). Additionally, we had previously recovered POULTRY_MG from wild asymptomatic finches (Hochachka et al. 2013), while Ferguson et al. (2003) had found HOFI_MG in turkeys. For these reasons, we decided to compare the bacterial load in choanal and conjunctival samples following experimental inoculation of House Finches with diverse POULTRY_MG and HOFI_MG strains.

Our experiments tested the following hypotheses. Hypothesis 1: M. gallisepticum changed tissue tropism following its evolutionary change to HOFI_MG. If that is correct, choanal swabs should have lower bacterial loads in birds inoculated with HOFI_MG than in birds inoculated with POULTRY_MG and higher bacterial loads in the conjunctiva of the birds inoculated with HOFI_MG. Hypothesis 2: M. gallisepticum expanded its within-host niche. This hypothesis predicts a similar bacterial load in the choana of birds infected with POULTRY_MG and HOFI_MG; the conjunctival load, however, should be higher in birds inoculated with HOFI_MG than in birds inoculated with POULTRY_MG.

In February and March 2019, and again in the fall of 2020, juvenile House Finches were trapped in and around Ithaca, New York (42°46′N, 76°45′W). Trapping activities were conducted under New York State Fish and Wildlife license 39 (Albany, New York) and permit 22669 from the United States Geological Survey, Department of the Interior (Laurel, Maryland). To test if the birds had been previously exposed to M. gallisepticum, we used a blood sample taken from each bird's brachial vein in a rapid plate agglutination (RPA) assay to test for the presence of M. gallisepticum–specific antibodies as described (Sydenstricker et al. 2006). Birds negative for antibodies were transferred to aviaries in a large barn (Dhondt et al. 2012), housed individually in wire bar cages (45 × 45 × 75 cm), and consumed food and water ad libitum. The food consisted of two-thirds pelleted diet (Daily Maintenance Diet, Roudybush Inc., Woodland, California, USA) and one-third black sunflower seeds (Wild Birds Unlimited, Indianapolis, Indiana, USA). Ceramic heat lamps were used as needed to prevent the water dishes from freezing. The lights were on an automatic timer and provided the finches 12 h of daylight each day. The experimental activities were approved by Cornell University's Institutional Animal Care and Use Committee (protocol 2009-0034).

After a minimum of 1 wk of quarantine, birds were inspected for the presence of eye lesions and retested for antibodies and for the presence of M. gallisepticum DNA. Only birds that remained negative were included in the experiment.

On 4 April 2019 (day 0), three groups of juvenile House Finches (n=6 or 7) were inoculated with 50 µL of inoculum in each eye. The strains used had been grown in the Geary Lab (University of Connecticut, Storrs, Connecticut, USA) and were used at the concentration provided. The inocula used are listed in Table 1. The F and the R_low strain are standard laboratory strains; the NY2001 strain was obtained from an asymptomatic House Finch (Ithaca, New York, USA) in 2001 (see Hochachka et al. 2013; Ley et al. 2016) but phylogenetically belongs to the POULTRY_MG group.

On 26 October 2020 (day 0), three groups were inoculated as in the previous experiment. The HOFI_MG strains used were CA2006 (n=5), CA2015 (n=4), and NC2006 (n= 5; Table 1). A negative control group (n=6) was sham inoculated with Frey medium provided by Edan Tulman (University of Connecticut, Storrs, Connecticut, USA). As all control birds remained negative throughout the experiment, we will not further report the results from these birds.

To determine whether inoculation led to infection, we examined the House Finches for conjunctival lesions, recorded the eye score (Sydenstricker et al. 2005), measured bacterial load via swab samples from the conjunctival sac and choanal cleft, and assessed the presence of M. gallisepticum–specific antibodies from blood samples. Before inoculation, baseline data on antibodies (from blood samples) and mycoplasmal bacterial load (from conjunctival and choanal swabs) for all the birds in both experiments were collected. Following inoculation, eye lesions were recorded, and choanal and conjunctival swabs for quantitative PCR (qPCR) and blood samples for antibody testing were taken on days postinoculation (DPI), 5, 10, 15, 21, 28, in 2019 and on DPI, 4, 7, 10, 14, 21, 28, in 2020. Eye lesions were recorded on a scale from 0 (no lesions) to 3 (severe lesions) following Sydenstricker et al. (2006). The choanal and conjunctival samples were placed separately each in a tube with 300 µL of Frey medium and stored at –80 C. Blood samples for RPA assays were taken from the brachial vein and placed in microcapillary tubes. After centrifugation, the plasma was separated to be used in the antibody test (Kleven 2008). Upon termination of the study, the birds were humanely euthanized by means of carbon dioxide gas inhalation.

To assess M. gallisepticum loads, DNA was extracted from the conjunctival and choanal swab samples using a DNeasy 96 Blood & Tissue Kit (Qiagen, Valencia, California, USA) following the manufacturer's recommended protocol. The total DNA was used in quantifying the mycoplasmal bacterial load by means of qPCR, as described by Grodio et al. (2008) and modified by Hawley et al. (2013). The target gene of interest was mgc2, and the number of mgc2 copies per sample reflects the bacterial load (Hawley et al. 2013). Given that both House Finch and poultry isolates of M. gallisepticum were used in this experiment, two different probes were used in qPCR reactions to assess bacterial load. Samples from birds in groups inoculated with a POULTRY_MG isolate or sham inoculated with Frey medium were tested with a qPCR using the poultry MG probe (IDEXX, Westbrook, Maine, USA) specific for poultry M. gallisepticum. All birds inoculated with a HOFI_MG isolate were tested using the House Finch–specific M. gallisepticum qPCR probe described by Grodio et al. (2008), as modified by Hawley et al. (2013) and Fleming-Davies et al. (2018).

Bacterial loads were log₁₀ transformed for statistical analyses. To compare the results of the bacterial loads between the two experiments, we calculated the area under the curve (AUC) between DPI 4 and 28. Because in 2019 the first sampling day was DPI 5, we calculated the AUC between DPI 5 and 28, calculated the mean bacterial load value per day, and added that to the DPI 5-28 AUC value to have comparable values. For each individual, we summed the eye scores for both eyes recorded on DPI 4 (5), 10, 14 (15), 21, and 28 (the DPIs on which eye scores were done in the 2020 experiment only are in brackets). We used analysis of variance (ANOVA) to compare the bacterial loads and eye scores between groups of birds inoculated with different isolates. If the ANOVA yielded a statistically significant result, we performed a Tukey honestly significant difference (HSD) post hoc test to make all pairwise comparisons. To compare how choanal and conjunctival loads differed from one another, we performed a one-sample t-test. All statistical tests were performed using Statistix 10 (Analytical Software 2021).

All House Finches were successfully infected, as shown by all developing M. gallisepticum–specific antibodies. Antibodies were first detected in 14 birds inoculated with HOFI_MG by DPI 4, while in 18 of 19 birds inoculated with POULTRY_MG, antibodies were first detected only by DPI 10 (χ²=29.2; df=1; P<0.0001), indicating a significantly more rapid immunologic response to infection when exposed to HOFI_MG.

None of the birds inoculated with the POULTRY_MG isolates developed eye lesions, while all birds inoculated with HOFI_MG had eye lesions (χ²=33.0; df=1; P<0.0001). The severity of the eye lesions summed to DPI 28 was significantly lower in the birds inoculated with CA2006 (mean 11.80±SE 1.202; A) compared with CA2015 (mean 27.67±1.552; B), and NC2006 (mean 29.80±1.202; B), as shown by an ANOVA (F₂,₁₂=58.43; P<0.0001) and a Tukey HSD post hoc test showing two homogenous groups: CA2006 A; and CA2015 B and NC2006 B.

Choanal M. gallisepticum loads did not differ significantly between isolates, as shown by an ANOVA (F_5,32=1.05; P=0.40). Conjunctival MG-loads, however, did differ between the isolates (ANOVA: F_5,31=55.98; P<0.0001; see Supplementary Material Table S1). The Tukey HSD post hoc test (at the α=0.05 level) indicated how the conjunctival bacterial loads differed between isolates: the three POULTRY_MG isolates formed one homogenous group with very low bacterial loads in the conjunctiva; and birds inoculated with HOFI_MG isolates had much higher bacterial loads. Among those, NC2006 demonstrated the highest load, and CA2006 had a significantly lower load. The CA2015 strain demonstrated statistically intermediate conjunctival loads (Table 2 and Fig. 1). When exploring the trajectories of the M. gallisepticum loads (Fig. 2), it can be seen that conjunctival loads of POULTRY_MG isolates remained very low throughout the experiment, while the loads of birds infected with HOFI_MG isolates remained high. The loads in birds infected with the low virulence CA2006 isolate decreased much more rapidly compared with those in birds infected with the more virulent CA2015 and NC2006 isolates. In contrast, the choanal loads of all isolates remain at a similar level throughout the experiment, although all declined by DPI 28.

The one-sample t-test showed that birds inoculated with the POULTRY_MG isolates had a much lower bacterial load in the conjunctiva than in the choana, while the opposite was true in birds inoculated with the virulent HOFI_MG strains (Table 2). In birds inoculated with CA2006, an isolate of low virulence, the bacterial loads did not differ between choanal and conjunctival swabs.

Our experiments clearly show that the conjunctival M. gallisepticum load in House Finches inoculated with a POULTRY_MG isolate was very low both compared with the choanal load in the same individual and compared with the conjunctival loads of birds inoculated with a HOFI_MG isolate. Choanal M. gallisepticum loads, on the other hand, did not differ between M. gallisepticum isolates, indicating that HOFI_MG has simply expanded its within-host niche to the eye as compared with its POULTRY_MG–like ancestors, in support of our hypothesis 2 and contrary to the predictions of hypothesis 1. Isolates of HOFI_MG cause a more rapid immunologic response in House Finches than POULTRY_MG and more severe lesions. Among the House Finch M. gallisepticum isolates, those that cause the most severe eye lesions also had the highest conjunctival load, as expected from Hawley et al. (2013), but no difference in the choanal load.

Motility is the capability of bacteria to reach and remain within host niches for nutrients or shelter from the host's defense mechanisms. Mycoplasma gallisepticum is known to have gliding motility, which appears to be essential for its parasitic lifestyle and for spreading in the respiratory tract once it has infected a host (Indikova et al. 2014). After the mycoplasmal conjunctivitis outbreak in wild birds, it was determined that strains of M. gallisepticum infecting House Finches form a separate phylogenetic clade (Hochachka et al. 2013). Tulman et al. (2012) reported a few coding changes of genes related to cytoadherence, such as hlp3, plpA, mgc2, gapA, crmA, and crmB-like. Other studies with different strains of M. gallisepticum have reported a high variation in cell adherence among HOFI_MG isolates (Perez et al. 2020), and compared with POULTRY_MG, HOFI_MG strains were shown to be qualitatively more adhesive and invasive than the poultry counterpart (Dowling et al. 2020), which seems to be related to coding changes, in particular, of three proteins MGC2, GapA and CrmA, essential not only for cytoadherence but also for gliding motility (Indikova et al. 2014). These reported observations align with hypothesis 2, because HOFI_MG isolates with a better adherence might be able to better spread within the host. Additionally, although Perez et al. (2020) showed a variation in adherence among HOFI_MG isolates, they did not find any sign of changes in glycan linkage specificity and binding avidity that are associated with changes in tissue tropism.

Sampling is a critical stage for diagnosis in live animals, especially for cases with low bacterial loads. This study has confirmed that when testing birds for infection with M. gallisepticum, it is important to sample and test the choana and the conjunctiva separately. Measuring bacterial loads in samples pooled from the two sites may be misleading.

We thank Adrianne Chissus for capturing the birds and helping with the experiments. We thank Avery August and Toshi Kawate, College of Veterinary Medicine at Cornell University, for use of their facilities. This work was supported by the Cornell University multistate grant 1020824, awarded through the NC1180 program Control of Emerging and Re-emerging Poultry Respiratory Diseases in the United States funded through the National Institute of Food and Agriculture – USDA.

Supplementary material for this article is online at http://dx.doi.org/10.7589/JWD-D-21-000187.