Analytically Sensitive Rickettsia Species Detection for Laboratory Diagnosis

INTRODUCTION

Rickettsia species (spp.) are Gram-negative obligate intracellular Alphaproteobacteria and the causative agents of rickettsioses throughout the world. Classified according to their antigenic properties, two main groups are associated with human disease: spotted fever group Rickettsia (SFGR), responsible for diseases such as Rocky Mountain spotted fever (RMSF), African tick bite fever, and Mediterranean spotted fever; and typhus group Rickettsia which cause murine and epidemic typhus. The transitional group of Rickettsia is responsible for diseases such as rickettsialpox and flea-borne spotted fever. These rickettsioses often present with undifferentiated symptoms during early illness, such as fever, headache, and malaise, making clinical diagnosis difficult, so cases are often missed even in endemic areas or when severe.^1,10,15 Certain SFGR present with rash or eschar (necrotic lesion), but these cannot be relied on for diagnosis. Epidemiologic information (e.g., travel, arthropod or animal exposure) and prompt clinical responses to doxycycline (the recommended treatment) also support clinical diagnosis. Early treatment is essential to increase the probability of a positive prognosis, especially in RMSF, for which a delay in doxycycline administration is correlated with more severe outcomes and death.¹⁰

The serological standard for confirmation of rickettsioses is the observation of a ≥ 4-fold rise in patient IgG antibody titers by indirect immunofluorescence antibody assay (IFA) to rickettsial antigens in paired acute (first 2 weeks of illness or while symptomatic) and convalescent samples taken 2 to 4 weeks later.^3,15 Recently, molecular detection by real-time polymerase chain reaction (PCR) of rickettsial DNA in clinical specimens has become more widely used for accurate specific confirmation of rickettsial infections in patients.^8,9 Although early treatment is beneficial, samples must be obtained before or within 48 hours of treatment or PCR sensitivity is reduced. Although eschar swabs and tissue samples may contain abundant organisms, rickettsial bacteremia is generally observed to be low, at < 100 genomic copies (gc)/mL in peripheral blood, except in advanced disease.^6,7,12 Even at the lower limits of detection (LoD) for real-time PCR, an efficiently repeatable assay at ∼9 DNA gc/reaction (5 µL of sample) still requires ∼1,800 gc of organism in 1 mL of blood.⁶ These factors are particularly troubling in the case of RMSF because of the rapid progression of this illness. While next generation sequence analyses of metagenomic samples are becoming more commonplace, most molecular assays target DNA of single copy genes in the rickettsial genome.⁸ We describe a real-time reverse transcriptase PCR (rtRT-PCR) assay, RCKr, for the detection of Rickettsia spp. including those known to be pathogenic to humans.

MATERIALS AND METHODS

The RCKr assay (PanR6 in U.S. Patent Application 2021/0079453) was developed to detect Rickettsia spp. targeting both 23S rDNA and rRNA in total nucleic acid (TNA) extractions. Primers and probe were designed using the Integrated DNA Technology OligoAnalyzer Tool (IDT; https://www.idtdna.com/pages/tools/oligoanalyzer). The rtRT-PCR assay was performed using 5 µL TNA and qScript XLT 1-Step RT-qPCR ToughMix, Low-ROX (Quantabio, Beverly, MA) with the RCKr primer/probe set (forward primer: GGTCCCACAGACTTACCAAACTCA, reverse primer: TCGACTATGGACCTTAGCACCCAT, and probe labeled with a fluorescein (Fl) and black hole quencher-1 (BHQ1): Fl-CCGAATGTCGATGAGTACAGCATAGCAGAC-BHQ1). The Applied Biosystems 7500 Fast Dx Real-time PCR Instrument (Thermo Fisher Scientific, Waltham, MA) was used for testing with the following cycling parameters: 50°C for 10 min, 95°C for 1 min, and 45 cycles of 95°C for 10 sec, and 60°C for 1 min. Ramp rates for all steps were set at 7500 fast conditions, and fluorescence data was acquired at the end of each annealing step. PanR8, an existing real-time PCR assay developed for Rickettsia species DNA detection, was used for comparison with the RCKr assay. PanR8 primer and probe sequences and testing parameters were used according to Kato et al.⁷

R. rickettsii AZ3 strain isolated and cultured at the Rickettsial Zoonoses Branch, Centers for Disease Control and Prevention (Atlanta, GA), was used to assess assay performance.^13,14 R. rickettsii organism from cell culture supernatant was suspended in sucrose phosphate glutamine (SPG) buffer and quantified using the PanR8 assay. EDTA whole blood was collected from a healthy donor (institutional review board [IRB] protocol number IRB00045947) through the Emory University Centers for Transfusion and Cellular Therapies (Atlanta, GA). The R. rickettsii suspensions were spiked into blood at 10-fold serial dilutions from 20,000 to 2 gc/mL. TNA was extracted using the MagNA Pure Compact Instrument and the MagNA Pure Compact Nucleic Acid Isolation Kit I (Roche Diagnostics, Indianapolis, IN) with an external lysis protocol. The extractions were performed with 280 µL of the MagNA Pure LC Lysis/Binding Buffer (Roche Diagnostics, Indianapolis, IN) and 20 µL of proteinase K (Sigma-Aldrich Inc., St. Louis, MO) added to 200 µL of the contrived blood sample, then incubated for 30 minutes at room temperature and eluted in 200 µL. The extracted TNA was evaluated using RCKr and PanR8 for reproducibility testing with nine total replicates.

Inclusivity for RCKr was assessed using nine Rickettsia species TNA extracted from culture grown organisms (Table 1). Exclusivity was evaluated for the RCKr assay by testing seven Rickettsia near neighbors, seven human tissues, nine environmental background bacteria, ten bacteria that cause diseases with similar symptoms, and two Proteus species that exhibit antigenic cross-reactivity with Rickettsia (Table 1). For accuracy, 30 TNA extracts from contrived blood specimens containing concentrations of R. rickettsii at 1,000, 100, 10, and 1 copy per reaction were evaluated. Ten additional blood specimens containing no organism were included as negative samples. The specimens were randomized and blinded by one operator, then extracted for TNA and tested by a second operator. Intra-assay reproducibility testing was performed using TNA extracted from 10 contrived blood specimens containing variable concentrations (1,000, 100, 10, and 1 copy per reaction) of R. rickettsii on the same run, in triplicate. Interassay reproducibility testing was performed using 1 TNA at 1 copy per reaction assessed over the course of 10 runs with multiple operators. To validate the use of the RCKr assay on diagnostic samples (IRB protocol number IRB7014), 49 banked clinical samples from 35 fatal and nonfatal cases of suspected rickettsioses were tested using RCKr and PanR8 assays (Table 2).

RESULTS AND DISCUSSION

Extracted TNA from contrived blood samples spiked with R. rickettsii at 10-fold serial dilutions from 20,000 to 2 gc/mL were tested to assess analytical sensitivity (Figure 1A). The LoD for RCKr was determined to be 20 gc/mL at 89% reproducibility (8/9 replicates). Originally described with an LoD of 1,800 gc/mL, PanR8, in this dilution series for comparison, detected 2,000 gc/mL with 67% reproducibility, representing a ∼100-fold increase in detection for RCKr. Inclusivity tests using TNA from culture grown organisms from nine Rickettsia species were positive with RCKr, whereas exclusivity assessment consisting of 35 TNAs tested negative (Table 1). High accuracy and precision were determined using contrived blood specimens, and results correlated as expected even with equivocal PanR8 samples at one copy per reaction (Table 1).

View Full Size

(A) Limit of detection (LoD) assessment using contrived samples of Rickettsia rickettsii organism spiked into blood at 10-fold dilutions of 20,000 to two genomic equivalents per 1 mL. Three sets of blood samples at each concentration were extracted for TNA. All extracts were tested with the RCKr (rtRT-PCR) and PanR8 (rtPCR) assays for a total nine replicates. The number of positive replicates for each data set is represented as a percentage (reproducibility). *Expected PanR8 assay LoD7. (B) Detectable limits of Rickettsia in acute blood of patients using the PanR8 and RCKr assays. The lowest detectable dilution by DNA detection was estimated based on quantification values and the dilution meeting parameters of the established PanR8 assay LoD at nine copies/5 µL. The Ct values are in parentheses for the last dilution where Rickettsia was detectable.

Citation: The American Journal of Tropical Medicine and Hygiene 106, 5; 10.4269/ajtmh.21-0757

• Download Figure
• Download figure as PowerPoint slide

Forty-nine banked clinical samples were assessed using both RCKr and PanR8. In 18 clinical samples (1–18, Table 2) from eight fatal RMSF cases which included blood, serum, cerebrospinal fluid, plasma, brain, spleen, kidney, lung, liver, and skin, RCKr detected rickettsial target sequences an average of 5.7 (range 0.54 to 9.41) Ct values lower than PanR8, whereas in 13 samples (19–31) from 11 nonfatal patients that included blood, plasma, bone marrow, swab, and eschar samples, RCKr detected rickettsial target nucleic acid at an average of 3.1 (range –1.3 to 6.12) Ct values lower than PanR8. Ct values of three specimens (27–29) were detected at similar values for both assays, and one specimen (31) had a Ct value difference of 1.3 higher with RCKr than PanR8 (Table 2). This could be due to nucleic acid degradation. Samples from patient 19 included swab, eschar, and blood (samples 30–32, Table 2). The blood, sample 32, collected on day 8 post–symptom onset (pso) and before doxycycline administration, was negative when originally tested with PanR8, and positive with RCKr. The negative PanR8 and high RCKr Ct values indicate undetectable or very low rickettsial copy numbers. The swab and eschar samples (30–31) from patient 19, collected 15 days pso and 4 days post–doxycycline treatment, were positive by PanR8 and RCKr.

Of the eight specimens (42–49, Table 2) collected from seven patients with unconfirmed rickettsioses, four specimens (three patients) were positive by PanR8 and RCKr, and one specimen (one patient) was positive by RCKr alone (amplicon size verified using gel electrophoresis). It remains unclear whether the RCKr positive and PanR8 negative patient sample represents an increase in Rickettsia detection or a false positive; although diagnosed clinically as rickettsiosis and with amplicon size verification, species identification through real-time PCR or sequencing attempts were unsuccessful (possibly due to low DNA quantities). Twelve specimens previously determined to be negative by PanR8 (33–41 and 47–49, Table 2) were also negative by RCKr. Of these 12, 10 etiologies remain undefined, whereas samples 38 and 39 previously tested positive by PCR for Ehrlichia chaffeensis and Anaplasma phagocytophilum, respectively. RCKr detected all previously defined positive samples as well as two additional specimens that PanR8 was not able to detect.

When comparing the performance of PanR8 and RCKr, lower Ct values indicate a higher number of detectable PCR target sequences and therefore the potential for better detection. The estimated increases in RCKr detectable targets were observed from 10-fold serial dilutions of Rickettsia positive sample extracts (2, 3, 19, 22-24, Table 2), which shows detection at 100 to 1,000,000-fold dilutions (Figure 1B). The variability in the difference between DNA and TNA detection probably reflects the presence and stability of rRNA in individual samples. It is unknown how this difference is affected by disease severity, day collected, sample age, or doxycycline treatment.

Because IFA is the reference standard for rickettsial disease confirmation, it would be valuable to assess acute samples collected from patients who demonstrated seroconversion (4-fold change in IFA titer) to better assess the performance of these assays. Although swabs, eschars, and tissue samples contain abundant rickettsial organisms during acute illness, blood and serum typically contain lower quantities due to low levels of circulating bacteria, which may not be measurable with PanR8.¹⁶ However, RCKr could provide more reproducible detection because of the lower Ct values and higher abundance of rRNA targets within a sample. This improvement may provide further insight into the early stages of disease in these sample types.

23S rRNA is constitutively expressed and bound by ribosomal proteins and is therefore potentially more stable and in higher copy number than the more traditionally designed DNA or mRNA targets. More 23S rRNA copies per Rickettsia cell are present than the single copy chromosomal DNA target. The 23S ribosomal RNA target region is in the operon that includes the methionyl-tRNA_f^Met formyltransferase (fmt), 23S rRNA, and 5S rRNA gene cluster, which are well conserved among Rickettsia including those that are pathogenic to humans. This gene arrangement is unlike the 16S (rrs), 23S, and 5S rRNA operon observed in many other bacteria. This uncommon arrangement seen in Rickettsia is thought to have occurred before Rickettsia spp. and Orientia spp. diverged.² This gene arrangement, and more specifically, the conserved sequences flanking the 23S rRNA and 5S rRNA intergenic region has previously served as a target for Rickettsia detection, whereas sequence variation in the intergenic region is used for differentiation of species.⁵ The 16S rRNA gene has also been used for Rickettsia detection because of the highly conserved sequences found among the species in this genus.^4,11

Relatively little is documented about the stability of ribosomes and rRNA in Rickettsia under different storage conditions. Although samples drawn after 48 hours of doxycycline administration may have reduced rickettsial content, we were able to detect rickettsial DNA and rRNA in specimens drawn after 72 hours (5–7 days pso) in patients with fatal outcomes. Swabs and eschar samples from nonfatal rickettsiosis patients remained positive at 15 or more days pso. The swab and eschar (patient 19) collected 15 days pso were sampled 96 hours post doxycycline. These results show detectable Rickettsia may be present in samples such as swab and eschar outside of the recommended collection timeframe (e.g., within 14 days pso and before or within 48 hours of doxycycline treatment), adding to the significance of collecting specimens while the patient is symptomatic and not restricting sample collection to a general timeframe of 14 days.

CONCLUSION

We show an increase in detectable nucleic acid by 100-fold in clinical specimens with more reproducible Ct values when including 23S rRNA as a target, therefore improving the accuracy of rickettsial detection. Future work includes evaluation of additional samples throughout the course of infection to determine the utility of this protocol in other rickettsial species of interest and to better understand assay limitations for agent detection at the early stage of illness. Regardless, our results represent an important improvement in effective laboratory diagnosis of rickettsioses.