Isolation, biochemical and molecular identification of Nocardia species among TB suspects in northeastern, Tanzania; a forgotten or neglected threat?

Background Pulmonary nocardiosis mimic pulmonary tuberculosis in most clinical and radiological manifestations. In Tanzania, where tuberculosis is one of the major public health threat clinical impact of nocardiosis as the cause of the human disease remains unknown. The objective of the present study was to isolate and identify Nocardia isolates recovered from TB suspects in Northeastern, Tanzania by using biochemical and molecular methods. Methods The study involved 744 sputum samples collected from 372 TB suspects from four periphery diagnostic centers ﻿in﻿ Northeastern, Tanzania. Twenty patients were diagnosed as having presumptively Nocardia infections based on microscopic, cultural characteristics and biomèrieux ID 32C Yeast Identification system and confirmed using 16S rRNA and hsp65 gene specific primers for Nocardia species and sequencing. Results Biochemically, the majority of the isolates were N. asteroides (n = 8/20, 40%), N. brasiliensis (n = 4/20, 20%), N. farcinica (n = 3/20, 15%), N. nova (n = 1/20, 5%). Other aerobic actinomycetales included Streptomyces cyanescens (n = 2/20, 10%), Streptomyces griseus, Actinomadura madurae each (n = 1/20, 5%). Results of 16S rRNA and hsp65 sequencing were concordant in 15/17 (88. 2%) isolates and discordant in 2/17 (11.8%) isolates. Majority of the isolates belonged to N. cyriacigeorgica and N. farcinica, four (23.5%) each. Conclusions Our findings suggest that Nocardia species may be an important cause of pulmonary nocardiosis that is underdiagnosed or ignored. This underscores needs to consider pulmonary nocardiosis as a differential diagnosis when there is a failure of anti-TB therapy and as a possible cause of human infections. Electronic supplementary material The online version of this article (doi:10.1186/s12879-017-2520-8) contains supplementary material, which is available to authorized users.


Background
Members of genus Nocardia are characteristically grampositive, weakly acid-fast, strictly aerobic, filamentous branching bacilli that fragment into rod or coccoid shaped forms. Nocardia species are ubiquitous environmental bacteria capable of causing opportunistic infections in both human and animals [1,2]. Nocardia species are increasingly isolated as infectious agents in immunocompromised patients, and at times, even among healthy individuals [3], causing infections ranging from pulmonary, cutaneous and subcutaneous human diseases [4].
The incidence of nocardiosis varies geographically according to a number of factors, like the prevalence of HIV infections, transplants, cancer, climate as well as socio-economic status, and laboratory capacity for Nocardia species detection and identification.
Increased incidence of human nocardiosis may be attributed to the wide use of immunosuppressive drugs, improved diagnostic tests, and increased awareness among microbiologists and health professionals. Nonetheless, in many developing countries where other chronic lung diseases, particularly TB, are prevalent, Nocardia species are either missed or misidentified during diagnosis [9,10].
Accurate detection and identification of Nocardia species have become increasingly important for prediction of antimicrobial susceptibility since different species have emerged in terms of their virulence and epidemiology. Furthermore, prompt and timely species identification can significantly influence the choice of therapy.
Apart from Gram and modified acid-fast staining, identification of Nocardia species depends largely on biochemical tests and cellular fatty acid analysis, which have proved to be laborious with long turnaround time, and less definitive.
Several molecular identification methods have been successfully employed to identify and characterize Nocardia species. Multilocus sequence analysis (MLSA) of 16S rRNA, the 65-kDa heat shock protein (hsp65), gyrase B of the ß subunit of DNA topoisomerase (gyrB), subunit A of Sec A preprotein translocase (secA1) and RNA polymerase (rpoB) have previously been used to identify Nocardia species [11][12][13]. However, 16S rRNA gene sequencing, which is the most often used system of bacterial identification, cannot discriminate many species [14], therefore identification based on the basis of the DNA sequence of a single housekeeping gene is hindered by stochastic genetic variation as well as horizontal gene transfer and recombination [15].
Pulmonary nocardiosis mimics pulmonary tuberculosis in both clinical manifestations, being chronic in nature and radiological characteristics makes it difficult to differentiate from M. tuberculosis and may as well be often wrongly treated with anti-TB drugs [16,17]. Patients might also be confused with other chronic lung infections such as invasive fungal infection [18,19].
In Tanzania, where tuberculosis is one of the major public health threats and a third leading cause of adult morbidity and mortality after Malaria and HIV/AIDS; clinical reports of Nocardia are rare if found at all. Moreover, the knowledge, as well as the clinical impact of nocardiosis, is unknown, suggesting that infections due to this genus may be underdiagnosed and/or neglected as a cause of human diseases.
Due to the lack of information on pulmonary nocardiosis in Tanzania, the objective of the present study was to isolate and identify Nocardia isolates recovered from TB suspects in Northeastern, Tanzania by using biochemical and molecular methods.

Collection and decontamination of clinical specimens
Three hundred and seventy-two (n = 372) self-presented TB suspects were included in this study after presumptively diagnosed as pulmonary tuberculosis cases by clinical symptoms and microscopically by acid-fast Ziehl-Neelsen (ZN) and or fluorescence staining method from four peripheral diagnostic centers (PDCs) namely Makorora health centre, Ngamiani health centre, Bombo regional hospital and Muheza designated district hospital in Northeastern, Tanzania. From each patient, two sputum specimens were collected one spot during the first visit at the respective TB facilities and one early morning specimen. The specimens were decontaminated by Nacely L-cystiene (NALC) sodium hydroxide (NaOH) method as described previously [20]. Decontaminated specimens were concentrated by centrifugation at 3000×g for 20 min after discarding the supernatant the sediments were resuspended by adding 1.5 ml of phosphate buffered saline (PBS). All procedures were carried out in a certified level II biosafety cabinet.

Culture and identification of the isolates
Two drops of the centrifuged sediments were inoculated on Löwestein-Jensen (L-J) and incubated at 37°C for 8 weeks. Subsequently, 500 μl of each specimen was inoculated in a BacT/Alert bottle, incubated in the BacT/Alert® 3D system for 8 weeks. The sample was considered negative when no bacterial growth or positive signal from the BacT/Alert instrument was observed after 8 weeks incubation. Positive cultures were stained by ZN for the confirmation of the acid-fast bacilli (AFB). Confirmation of M. tuberculosis was performed by using GenoType® MTBC (Hain Life science GmbH, Nehren, Germany) and GenoType® Mycobacterium CM/AS assay (Hain Life science GmbH, Nehren, Germany) for the detection of common and accessory nontuberculous mycobacteria (NTM).
Negative cultures for AFB were cultured on blood agar (BA) and chocolate agar plates and incubated at 37°C for 2 to 4 days. Isolates were diagnosed as Nocardia based on the presence of non-acid-fast branching filamentous bacilli using Gram and modified Kinyoun staining, as well as their colony morphology.

Biochemical identification
The BioMèrieux ID 32C yeast identification system was used to identify the presumptively 20 Nocardia isolates diagnosed in this study. All procedures followed manufacturer's recommendations. Briefly, test isolates were grown on brain heart infusion agar or chocolate agar plates at 35°C. A Mc Farland no.4 was prepared in a standard physiological saline. One ml of the suspension was inoculated into ID 32C medium, and 3 drops were dispensed into each well of the strip using an automated dispenser ATB Vitek® 1574 (biomèrieux). The strips were incubated at 35°C for 7 days in a sealed container to avoid evaporation. The strips were read to give an eightdigit profile, as per manufacturer's instruction.

DNA extraction
DNA extraction was performed by using Ultraclean® Microbial DNA isolation kit (MOBIO Laboratories, Inc.) following manufacturer's instructions, after an initial heat inactivation of a loopful colony suspended in 500 μl-distilled water in a 2 ml Eppendorf tube then incubation at 95°C for 20 min followed by centrifugation at 15,000 x g for 5 min at room temperature (r.m.t). 50 μl of collected DNA was then stored at -20°C until use. M. tuberculosis reference strain H37Rv genomic DNA and distilled water were used as positive and negative controls respectively.

PCR for 16S rRNA
A 606-bp fragment of the 16S rRNA gene specific for Nocardia species was amplified with primers Noc1 (5′-GCTTAACACATGCAAGTCG-3′) (positions 46 to 64, Escherichia coli numbering system) and Noc2 (5′-GAATTCCAGTCTCCCCTG-3′) (positions 663 to 680, E. coli numbering system) [1]. PCR reaction was performed in a 25 μl volume with master mix contained a final concentration of 5 U Taq DNA polymerase (Applied Biosystems), 10× buffer, 10 mM dNTPs mix, 100 pmol/μl of each primer, 25 mM MgCl 2, and 100 ng template DNA. PCR amplification was carried out with an initial denaturation of 94°C for 5 min, followed by 40 cycles of denaturation at 94°C for 1 min, annealing at 58°C for 1 min, extension at 72°C. Then followed by a final extension at 72°C for 5 min. PCR products were purified with Qiagen PCR purification kit (Qiagen, CA, USA) following manufacturer's recommendations and were submitted for sequencing at the Department of Biochemistry, University of Leipzig, Germany.

PCR specific for hsp65 gene
A 440-bp fragment of the hsp65 gene encoding the 65 kDa heat shock protein was amplified using (TB11, 5′-ACCAACGATGGTGTGTCCAT-3′) and (TB12, 5′-CTTGTCGAACCGCATACCCT-3′) oligonucleotide primers [21]. PCR was performed in a final volume of 25 μl (5 U of Taq DNA polymerase (Applied Biosystems); 10× buffer, 25 mM MgCl2, 10 mM dNTPs mix, 100 pmol of each primer) with 10 μl of DNA template. Amplification was carried out in a thermal cycler (Applied Biosystems). Amplification cycles included an initial denaturation at 94°C for 5 min, followed by 35cycles (denaturation 94°C for 1 min, annealing at 55°C for 1 min, extension at 72°C for 1 min) and a final extension at 72°C for 10 min. For sequencing the PCR product of the hsp65 gene were purified by Qiagen PCR purification kit as recommended by the manufacturer (Qiagen, CA, USA) and submitted for sequencing at the Department of Biochemistry, University of Leipzig, Germany). Same primers used for PCR served for sequencing of the forward and reverse fragments.
Sequencing of 16S rRNA and hsp65 PCR product fragments PCR products were purified and sequenced on both strands. The resulting 16S rRNA and hsp65 sequences of the 20 isolates were analyzed using Codon Code Aligner software Version 5.1.5 (http://www.codoncode.com/ aligner/). 16S rRNA gene sequences were compared with the corresponding sequences of the Nocardia species in the GenBank database (www.ncbi.nlm.nih.gov) and LeBIBI QBPP database using the SSU-rDNA-16S_TSstringent [22] where there were disparities the LeBI-BI QBPP assigned species name was adopted; for the hsp65 gene only the GenBank database was used. For the comparison with GenBank database the Basic Local Alignment Sequencing Tool for nucleotide sequence queries (BLASTN) was used. The GenBank query type strain or culture collection strain with the highest score was downloaded and compared to the subject sequences and the percentage similarity was then determined for each strain. Strains with greater than 99% homology to a single type strain were considered definitive for identification to species level.

Ethical clearance
The protocol for this study was reviewed and approved by the Ethical Committee of the National Institute for Medical Research (NIMR) (Ref. No. NIMR/HQ/R.8a/ Vol.IX/1401), Dar es Salaam, Tanzania. Written informed consent was obtained from the patients or relative of the patients, where the patients could not read and write. Ethical approval was obtained for the secondary use of the study data from the same ethical committee.

Detection of Mycobacteria and Nocardia species
Of 372 decontaminated samples sixty-five (n = 65) were AFB positive by ZN staining, whereas one-hundred thirty-seven (n = 137) including the AFB positive were culture positive by either L-J or BacT/Alert 3D system. Two hundred and two (n = 202) samples were both AFB and culture negative and 33 samples were contaminated. A specimen was considered as contaminated when there was a growth of non-mycobacterial organisms in either L-J or BacT/Alert bottle. The purity of all fluid cultures by plating them on blood agar and chocolate agar plates and re-examining individual specimen by ZN-stain.
Eighty-one of the AFB isolates were confirmed to be M. tuberculosis using GenoType® MTBC (Hain Life science GmbH, Nehren, Germany) and 36 isolates were confirmed as NTM by GenoType® Mycobacterium CM/ AS (Hain Life science GmbH, Nehren, Germany). Twenty (5.4%) out of 372 TB suspects which are subject of this manuscript were presumptively diagnosed as having Nocardia infection. Of these nine (9/20) isolates were among the 137 isolates mistakenly recorded as AFB positive at the PDCs (see Additional file 1). Diagnosis of Nocardia species based on the presence of nonacid-fast branching filamentous bacilli using Gram and modified ZN staining, as well as their colony morphology on LJ, BA and Chocolate agar (Fig. 1). Eleven isolates grew well on LJ media, and all isolates grew on both BA and Chocolate agar.

Identification of Nocardia species by 16S rRNA and hsp65 gene sequencing
Since Streptomyces species are difficult to identify by the biomèrieux ID 32C Yeast Identification system because they give a less distinct profile pattern than Nocardia isolates [23], all 20 isolates tested by this method for the 16S rRNA were analysed using the LeBIBI QBPP a web tool [22] and the GenBank database, where there was disparities between the two databases the species name assigned by the former was adopted, and for sequences without the closest proximal cluster in LeBIBI QBPP , the closest sequence was identified based on the tree determination. Sequencing of hsp65 gene revealed that all 17 confirmed Nocardia isolates had closest species similarity identity score of >99% with type or culture collection strain in the database on one hand, and on the other the 16S rRNA sequencing, thirteen (n = 13) isolates had closest species belonging to the proximal clusters in the LeBIBI QBPP database with four (n = 4) isolates not belonging to the closest proximal cluster, hence their identification was based on the tree determination. Generally, hsp65 and 16S rRNA gene sequencing were concordant in 15 out of 17 (88.2%) isolates and discordant in two out of 17 (11.8%) isolates. Of the discordant results 16S rRNA identified one isolate as N. jinanensis TDQ462650 proximal cluster and another as N. thailandica TGQ376186 proximal cluster, whereas hsp65 identified these isolates as N. testacea and N. cyriacigeorgica respectively ( Table 2). Details of the isolates and their accession numbers are provided in Additional file 1.

Demographic and clinical data of patients diagnosed with Nocardia infection
Based on the 16S rRNA and hsp65 sequencing results the confirmed 17 cases of Nocardia isolates from this study, we revised the demographic and clinical data of all the patients obtained during the study period as shown in Table 3. The clinical symptoms and concurrent conditions show that the proportion of Nocardia infection was higher (58.8%) in individuals with ≥35 years of age than among those ≤35 years of age (41.2%), and both sexes are equally affected. The data further showed that nearly all patients presented with clinical syndromes mimicking those of pulmonary tuberculosis. The proportion of individuals with Nocardia infection alone was 76.5% (n = 13), three (n = 3, 17.7%) cases had coinfection with HIV and two (n = 2, 11.8%) cases were co-infected with TB and they had commenced anti-TB Fig. 1 Growth of Nocardia spp. isolated from TB suspects on three different media used in this study (a) LJ medium (see arrows), (b) Blood agar plate and (c) Chocolate agar plate treatment. Ten (n = 10, 58.8%) other cases were initiated with broad-spectrum antibiotic treatment whereas, for five (n = 5, 29.4%) cases records showed that they had no treatment whatsoever.

Discussion
Nocardia infections causing both human and animal diseases are increasingly reported owing to improved diagnostics especially in developed settings which have emphasized the need for rapid characterization of clinically isolated Nocardia [4,8,10,24,25]. However, data regarding nocardiosis from resource-poor settings like Tanzania heavily stricken by HIV and TB pandemic are rare if available at all. Such diseases are either underdiagnosed or neglected due to the similarity of clinical and radiological features between pulmonary nocardiosis and pulmonary tuberculosis on one hand and on the other due to poor diagnostic capabilities. In Tanzania where TB is still a major public health threat, anti-TB treatment is initiated based on clinical symptoms, direct smear microscopy and radiological diagnosis where available.
In this study, all 372 TB suspects self-presented at four health facilities were diagnosed based on clinical symptoms and direct smear microscopy by acid-fast staining method at the respective clinics. Eighty-one (n = 81, 21.8%) patients were positive for M. tuberculosis by culture and confirmed by GenoType® MTBC and 36 (9.7%) patients had NTM infections confirmed by Geno-Type® Mycobacterium CM/AS. 202/372 patients sputum samples were negative by culture and 33 other samples were contaminated.
Twenty (5.4%) isolates out of 372 TB suspects were presumptively diagnosed as Nocardia species based on their colony morphology on different culture media and by microscopic appearance using modified ZN for weak acid-fast bacilli.
Isolates of Nocardia species in this study showed good growth on LJ, BA and chocolate agar (Fig. 1). Decontamination of sputum samples with NALC-NaOH in this study did not seem to affect the growth of Nocardia species as observed in a previous study [26].
Samples from 17 suspects were presumptively diagnosed as having Nocardia infection and three patients had infection with other Actinomycetales. Findings that majority of Nocardia species grow well on LJ in this study correspond with findings of other authors [25,27]. In this scenario, such growth may be confused with that of mycobacteria species, making the diagnosis more complicated. Furthermore, careful microscopic analysis should be thoroughly performed, since these two genera may present difficulties for an inexperienced microscopist [24]. As observed in the previous study [25], where Nocardia species were detected by the modified ZNmethod for weak acid-fast bacteria, we also detected Nocardia species by the ZN-method employed for mycobacteria in this study.
All the Nocardia species identified in this study have been reported to be associated with pulmonary nocardiosis in many parts of the world [24,[26][27][28][29].
The finding that the biomèrieux ID 32C yeast identification system identified correctly five Nocardia isolates to species level and two isolates to least genus level as Streptomyces (Table 1) when compared to 16S rRNA and hsp65 sequencing results is in agreement with those in a study conducted in Brazil [13]. In the Brazilian study, seven isolates were identified by 31 different phenotypic tests using six identification systems with only two isolates correctly identified by the phenotypic method, compared to multilocus sequence analysis (MLSA) results. However, the biomèrieux ID 32C yeast identification system used in this study appears to be useful in identifying Nocardia species and other aerobic actinomycetes as previously determined [23].
The discrepancies between the ID 32C yeast identification system and sequencing results suggest that a careful interpretation of results indicating different Nocardia species and other Actinomycetales is needed since such methods are not accurate as shown in Table 2. Moreover, although the ID32C system has been used for identification of many aerobic actinomycetes, it does not, however, give consistent results within a species. There is a greater potential for errors than with 16S rRNA or with hsp65 gene sequencing. For example ID32C system present a lot of difficulties in identification of Nocardia asteroides complex, this method uses eight-digit profile as those used for identification of yeast, whereas the last four-digits show a consistent correlation between isolates within species, the first four-digits show poor correlation within a species and between species. There is no single profile for each particular isolate and species with this method, hence this variation. Therefore, results of 16S rRNA and hsp65 gene sequencing are more reliable than those of the ID32C system. Therefore, molecular identification is necessary for definitive identification of Nocardia species [10,[30][31][32][33]. Important to note is that the following species: N. cyriacigeorgica, N. farcinica, N. brevicatena and N. nova identified by both 16S rRNA and hsp65 sequencing in this study were separated from the originally referred Nocardia asteroides and later found to be a group of bacteria with a heterozygous pattern of antimicrobial drug susceptibilities [34]. N asteroides complex was further separated and reorganized into different species on the basis of drug susceptibility patterns: Nocardia abscessus, Nocardia brevicatena-paucivorans complex, Nocardia nova complex (includes N nova, Nocardia veterana, Nocardia africana, Nocardia kruczakiae), Nocardia transvalensis complex, Nocardia farcinica, and N asteroides [5]. Moreover, Nocardia cyriacigeorgica, which constitute the majority of Nocardia species in this study (n = 4, 23.5%) was also differentiated from N asteroides and is becoming a more frequently identified clinically significant pathogen [35].
This separation may explain the observed discrepancy between the two methods. While phenotypic identification leads to misidentification, molecular identification can improve the diagnostic accuracy since some molecular targets can present high sequence similarity [2].
Looking into the demographic data of all 17 Nocardia cases in this study, it was observed that nine (52.9%) were males and eight (47.1%) were females. Seven (41.1%) were ≤35 years and ten (58.8%) were ≥35 years. This incidence suggests that nocardiosis can occur in both sexes and in different age groups with more or less the same frequency. However, of importance is that nearly all cases of nocardiosis identified in this study reported similar clinical symptoms as those of TB suspects with chronic coughing featuring in all cases (Table 3). This is not surprising since pulmonary nocardiosis is the most common clinical presentation acquired primarily by inhalation. The onset of symptoms can be subacute to more chronic and can include productive or a nonproductive cough, shortness of breath, chest pain, hemoptysis, fever, night sweats, weight loss, and progressive fatigue [4].   In this study, 12 patients had Nocardia alone; two had co-infection with TB and were AFB positive, whereas three had co-infection with HIV. Similar findings have been reported in other studies [8,36]. It is important that irrespective of a patient's immunologic status, the isolation of Nocardia from the respiratory tract or another body source should not be regarded as a contaminant or commensal organism [4].

Conclusion
In conclusion, our study reveals that Nocardia species are important causes of pulmonary disease and that may have been underdiagnosed and/or ignored altogether. It is also clear that mycobacterial infections and nocardiosis treatment differ, correct identification of the causative agent is, therefore, critical to avoid treatment failure. Nocardia species identified in this study, underscore the need to consider pulmonary nocardiosis as differential diagnosis especially when there is treatment failure with the standard anti-TB therapy, and as a possible cause of human infections.