An epidemic CC1-MRSA-IV clone yields false-negative test results in molecular MRSA identification assays: a note of caution, Austria, Germany, Ireland, 2020

We investigated why a clinical meticillin-resistant Staphylococcus aureus (MRSA) isolate yielded false-negative results with some commercial PCR tests for MRSA detection. We found that an epidemic European CC1-MRSA-IV clone generally exhibits this behaviour. The failure of the assays was attributable to a large insertion in the orfX/SCCmec integration site. To ensure the reliability of molecular MRSA tests, it is vital to monitor emergence of new SCCmec types and junction sites.

We investigated why a clinical meticillin-resistant Staphylococcus aureus (MRSA) isolate yielded false-negative results with some commercial PCR tests for MRSA detection. We found that an epidemic European CC1-MRSA-IV clone generally exhibits this behaviour. The failure of the assays was attributable to a large insertion in the orfX/SCCmec integration site. To ensure the reliability of molecular MRSA tests, it is vital to monitor emergence of new SCCmec types and junction sites.
We investigated why a clinical meticillin-resistant Staphylococcus aureus (MRSA) isolate, collected in Austria in late 2019, yielded false-negative results with two widely used commercial orfX/SCCmec junction assays (Cepheid GeneXpert MRSA/SA BC, BD MAX Staph SR). The aim of this study was to investigate false-negative results with these two assays. Therefore, we tested and sequenced the index isolate and other isolates assigned by microarray to the same strain (i.e. the European CC1-MRSA-IV [1]).

Index case
A 62-year-old patient with metastasised cancer was admitted with suspected pneumonia to the Medical University Hospital of Graz, Austria. Blood cultures (BACTEC, Becton Dickinson, Heidelberg, Germany) became rapidly positive and Staphylococcus aureus was identified by in situ hybridisation (PNA FISH, AdvanDx, Woburn, United States (US)). In order to identify meticillin-resistant S. aureus (MRSA), the blood culture was investigated using the GeneXpert MRSA/SA BC PCR (Cepheid, Sunnyvale, US). Simultaneously, rapid antimicrobial susceptibility testing (RAST) was performed [2,3]. The GeneXpert test was negative for MRSA but RAST revealed cefoxitin resistance after 6 h. Antimicrobial susceptibility testing (Vitek2, bio-Mérieux, Marcy-l'Étoile, France) confirmed meticillin resistance. BD MAX StaphSR (Becton Dickinson) yielded a negative MRSA result. Microarray-based characterisation (S. aureus Genotyping Kit 2.0, Abbott, Alere Technologies, Jena, Germany) detected mecA and assigned the isolate (Graz_511421- 19) to clonal complex CC1-MRSA-IV. Although the antibiotic treatment was adapted, the patient died shortly after because of tumour progression.

Isolates
Ten CC1-MRSA-IV isolates originated from the Sfanta Parascheva Hospital, Iasi, in north-eastern Romania [4]. Four isolates originated from the Dresden University Hospital, Saxony, Germany. Nineteen isolates from Regensburg University Medical Centre in Bavaria, Germany and five isolates from other Bavarian hospitals were also included. All included isolates had been collected and preliminarily analysed as part of earlier collaborations. Four fully sequenced reference strains were used as controls (Table 1). Additional controls comprised four isolates from Dresden that belonged to local epidemic strains or to another CC1-MRSA strain (isolate Dresden-220663).

Commercial MRSA assays
Test results are provided in Table 1. The index isolate tested negative using the BD MAX Staph SR assay (Lot 9303156). Isolates from Iasi and Dresden and controls were tested twice, in Graz and Dresden, using this assay (Graz, Lot 9303156; Dresden, Lot K55928980720210312). It failed to identify these 15 isolates although controls handled in parallel were correctly identified.
Testing of the index isolate with GeneXpert MRSA/SA BC (Lots 1000148707 and 1000179462) yielded negative results, too. Further investigations on this assay were not possible because that laboratory became involved in diagnostics for the coronavirus (Covid- 19) pandemic.

Genotyping by microarray and sequencing
All isolates were genotyped using the S. aureus Genotyping Kit 2.0, a microarray covering 333 different target sequences corresponding to ca 170 different genes. Target genes, assay protocols and sequences of probes and primers have been published previously [5]. Isolates were assigned to clonal complexes, strains and SCCmec types based on microarray data as described [5].
All isolates underwent whole-genome sequencing. DNA was extracted as for array experiments. Its quality was assessed as previously described [6]. The Nextera DNA Flex Library Preparation Kit (Illumina, Eindhoven, the Netherlands) was used and libraries underwent pairedend sequencing using the 500-cycle MiSeq Reagent Kit v2 (Illumina). Libraries were scaled to exhibit at least 50-fold coverage. Sequencing run quality was assured following cluster density and Q30 assessment. Raw sequence reads were trimmed using fastp 0.19.11 [7] and assembled using SPAdes v3.9.1 [8]. Contigs under 1,000 bp were removed.
The sequence of the SCCmec element one representative isolate, Iasi-95037, was deposited in GenBank (accession number: MT380478).

Description of the strain and its SCCmec element
Microarray profiling and genome sequencing showed that the index isolate belonged to a CC1-MRSA-IV clone previously described as 'European CC1-MRSA-IV' that may have emerged in south-eastern Europe [1,4,9]. A putative, meticillin-susceptible ancestor is common in Romania where this MRSA clone frequently observed already several years ago [1,4]. A high prevalence or outbreaks have been reported from Ireland [1], Italy [10] and Germany (North Rhine-Westphalia and Bavaria) [1,11]. In Regensburg, retrospective microarray-based typing of 3,067 isolates revealed that the occurrence of the European CC1-MRSA-IV clone increased from < 1% of typed MRSA between 2010 and 2013 to 9.4% in 2019. In Dresden, this strain has only sporadically been observed, accounting for seven in 1,758 isolates genotyped since 2000 ( [1,12] and data not shown). Microarray genotyping data indicated that this clone was also recovered from horses and wild birds in Austria [13,14] and from livestock in Italy [15].
The clone is PVL-negative and lacks the splE protease gene. It only rarely carries enterotoxin genes sek/ seq in addition to seh that is ubiquitously present in CC1. Its relationship to other CC1-MRSA clones has been discussed previously in detail [1].
An identical 5,350 nt cluster is present in Staphylococcus epidermidis P11PPP12 (MH188467.1). Beyond that, the entire SCCmec IV element in P11PPP12 is identical to the one in the CC1-MRSA strain. Significantly, the site of recombination cutting short Q7A213 is conserved in both strains (position 5,818/5,819 in the Supplement). Thus, it is likely that the entire SCCmec IVa cassette including the insert was transferred between ancestors of the two strains, i.e. across the species. However, P11PPP12 also harbours an ACME-II/heavy metal resistance element downstream of SCCmec, which is absent from the European CC1-MRSA-IV. Therefore, it must have been lost during or after transfer of the SCCmec element, or it was acquired later by S. epidermidis.

Discussion
The study demonstrates that a CC1-MRSA-IV epidemic strain in Europe can yield false-negative results with common MRSA assays (GenXpert MRSA/SA BC, BD MAX Staph SR). Interestingly, GeneXpert MRSA/SA SSTI yielded correct results, indicating that the different tests utilise different primers. The absence of dcs and coverage of SCCterm 15 appear to be the reason for the discrepancy.
False-negative results of PCRs targeting the orfX/ SCCmec junction site are concerning. Molecular assays are used to predict MRSA in positive blood cultures and to change therapy accordingly. The use of molecular assays is beneficial for a vast majority of patients because a result is available quickly. However, these assays can only detect target sequences that were available and considered at the time the primers were designed, and false-negative results have the potential  to harm the patient by delaying effective therapy. Conventional antibiotic susceptibility tests are slower but are not constrained by the choice of primers or by the presence of unknown genotypes.
Molecular assays are also used to guide infection control. False-negative tests may result in lapses facilitating further MRSA transmission. Another, less obvious consequence might be a shift in the clonal structure of MRSA populations. When molecular assays exert a selective pressure favouring a false-negative strain, PCR-positive strains might get 'penalised' with subsequent interventions, hindering proliferation and transmission. This could lead to an increasing prevalence of the false-negative strain and to more failures in therapy and infection control.
It is crucial to monitor the emergence of new SCCmec junction sites in S. aureus and in coagulasenegative staphylococci, as mobile SCCmec elements can readily be transmitted between different strains and species, as was the case in the strain described here. Such unknown genotypes represent a problem for established molecular assays. As illustrated here, updating existing tests and platforms to evolving genotypes of the target organisms is important for individual and public health.
The containment of the CC1 strain must rely on conventional susceptibility tests, culture-based screening using selective growth media or updated molecular tests. We propose screening of medical or nursing staff recruited from epidemic regions, not only in hospitals but also in other care facilities, as well as patients with travel histories to these regions.