Prevalence of Carbapenem-Resistant Hypervirulent Klebsiella pneumoniae and Hypervirulent Carbapenem-Resistant Klebsiella pneumoniae in China Determined via Mouse Lethality Tests

Objective To investigate the epidemiology of carbapenem-resistant hypervirulent Klebsiella pneumoniae (CR-HvKP) and hypervirulent carbapenem-resistant Klebsiella pneumoniae (Hv-CRKP). Methods Totally 436 K. pneumoniae strains were collected from 7 hospitals in mainland China between 2017.01 and 2018.02. Sequence types, serotypes, antimicrobial-resistance and virulence genes were analyzed. Additionally, string test, capsule stain, Periodic Acid Schiff stain, fitness analysis, quantitative real-time PCR and mouse lethality test were also performed. Molecular combinations were used to screen putative bla KPC(+)-HvKP and Hv-bla KPC(+)-KP, followed by the confirmation of mouse lethality test. Results Diverse detection rates were found for the virulence genes, ranging from c-rmpA (0.0%) to entB (100.0%). According to the molecular criteria, 127, 186, 9 and 26 strains were putatively denoted as HvKP, bla KPC(+)-KP, bla KPC(+)-HvKP and Hv-bla KPC(+)-KP. Mouse lethality test confirmed 2 bla KPC(+)-HvKP strains (JS184 and TZ20) and no Hv-bla KPC(+)-KP. JS184 showed K2 serotype, thin capsule, positive exopolysaccharid and string test. TZ20 presented K20 serotype, thin capsule, negative exopolysaccharide and string test. Compared with the positive control NTUH-K2044, equal galF expression and growth curves were confirmed for JS184 and TZ20. Conclusions Molecular determination of CR-HvKP and Hv-CRKP brings remarkable bias compared with mouse lethality test. The exact prevalence of CR-HvKP is less than 1.0%, which of Hv-CRKP is much lower.

Hypervirulence in K. pneumoniae represents another major concern. Hypervirulent K. pneumoniae (HvKP) was first reported to cause pyogenic liver abscess (PLA) and septic endophthalmitis in seven healthy individuals (Liu et al., 1986). HvKP, which has a considerably lower median lethal dose (LD 50 ) than that of cKP in mouse model, generally produces various virulence factors such as hypercapsules, excessive siderophores, exopolysaccharides, and fimbriae (Paczosa and Mecsas, 2016;Russo and Marr, 2019). Apart from PLA, HvKP can also cause multiple invasive infectious diseases such as endogenous endophthalmitis, necrotising fasciitis, and meningitis, and the infection can undergo metastatic spread. PLA is endemic to East Asia and associated with a morbidity rate of 15.45 per 100,000 person-years in 2011 and a mortality rate of 8.2% (Chen et al., 2016;Siu et al., 2012). It has been estimated that 60% of endogenous endophthalmitis cases are associated with PLA caused by K. pneumoniae (Wong et al., 2000). Even with intravenous and intravitreal antimicrobial treatment, 89% of endophthalmitis cases show visual acuity of light perception or worse, and over 40% of affected eyes require evisceration or enucleation (Yang et al., 2007).
Recently, a combination of hypervirulence and extreme drug resistance has been reported in K. pneumoniae, thereby exacerbating the scarcity of effective treatments and resulting in high mortality (Zhang et al., 2015;Gu et al., 2018). The prevalence of infections caused by carbapenem-resistant hypervirulent K. pneumoniae (CR-HvKP) and hypervirulent carbapenem-resistant K. pneumoniae (Hv-CRKP) presents a global concern and a great challenge in clinical practice. However, the epidemiology of CR-HvKP and Hv-CRKP has not been extensively studied. The estimated prevalence of CR-HvKP/Hv-CRKP strains ranges among 5.0-15.0% among CRKP strains in mainland China, based on molecular determination or the Galleria mellonella (greater wax moth) lethality test (Zhang et al., 2020;Zhan et al., 2017). The gold standard method for evaluating the virulence of K. pneumoniae involves the use of mouse models, rather than the G. mellonella lethality test (Russo and MacDonald, 2020). Here, we analysed 436 clinical K. pneumoniae strains using molecular techniques and mouse lethality tests to elucidate the prevalence of CR-HvKP and Hv-CRKP strains.

K. pneumoniae Strains
In this study, 436 non-duplicate and consecutive K. pneumoniae isolates were collected from seven hospitals across six provinces in China (Huashan Hospital, 180 strains; Jinshan Hospital, 28 strains; Taizhou Municipal Hospital, 84 strains; The First Affiliated Hospital of Guangxi Medical University, 20 strains; Kunming Yan'an Hospital, 34 strains; Sixth Hospital of Shanxi Medical University, 60 strains; Shandong Provincial Hospital Affiliated to Shandong University, 30 strains) from January 2017 to February 2018. The 436 isolates were collected from diverse sources: 255 isolates (58.5%) were obtained from sputum samples, 98 isolates (22.5%) from urine samples, 29 isolates (6.7%) from blood samples, and 54 isolates (12.4%) were obtained from other sources. All the isolates were cultured on sheep blood agar plates and kept at -80°C prior to use. Identification of K. pneumoniae was performed using a matrixassisted laser desorption/ionization time-of-flight mass spectrometry system (Bruker Daltonics Inc., Fremont, CA, USA) using the standard strains P. aeruginosa ATCC 27853, K. pneumoniae ATCC 700603, and E. coli ATCC 25922 as controls.
All the K. pneumoniae strains were investigated as the flow chart in Figure 1.

Determination of Serotypes, Antimicrobial-Resistance, and Virulence Genes
The capsule type was determined via PCR amplification and sequencing of the wzi loci (Brisse et al., 2013), followed by comparison with sequences on the database of Institute Pasteur (https://bigsdb.pasteur.fr/klebsiella/klebsiella.html).

String Test
Overnight cultured K. pneumoniae colonies on sheep blood agar plates were stretched outward using an inoculation loop as described previously (Shon et al., 2013). The string test was considered positive when a viscous string produced was over 5 mm in length. Strain NTUH-K2044 was used as a positive control and HS11286 was used as a negative control.

Capsule Staining
Capsule staining of K. pneumoniae strains was performed according to the manufacturer's instructions (catalog number: BA-4039; BASO, Zhuhai, China). NTUH-K2044 was used as a positive control and HS11286 was used as a negative control.

Periodic Acid-Schiff Staining
Periodic acid-Schiff staining was performed according to the manufacturer's protocol (catalog number: BA4080A; BASO, Zhuhai, China). Strains NTUH-K2044 and HS11286 were used as positive and negative controls, respectively.

Fitness Analysis
A growth curve was generated to evaluate the fitness of K. pneumoniae strains (Liu et al., 2016). These strains were cultured overnight in Luria-Bertani broth, diluted to an optical density at 600 nm (OD 600 ) of 0.001, and cultured at 37°C under aerobic conditions (BioTek Synergy H1, Winooski, VT, USA). The OD 600 values were measured every 30 min and plotted as a curve. Strains NTUH-K2044 and HS11286 were used as positive and negative controls, respectively.

Quantitative PCR Analysis
Quantitative PCR analysis of galF mRNA together with 16S rRNA was performed using an Applied Biosystems 7500 system (Applied Biosystems, San Ramon, CA, USA). The primers used are shown in Table S1. Strains NTUH-K2044 and HS11286 were used as positive and negative controls, respectively. The analyses were performed according to the manufacturer's protocol (catalog number: FS-Q1002; FOREVER STAR, Beijing, China).

Mouse Lethality Test
Mouse experiments were approved by the Institutional Animal Care and Use Committee of the School of Pharmacy, Fudan University (Shanghai, China) (ethical approval document NO. 201603-TY-MQ-01). Pathogen-free female BALB/c mice (age, 6 weeks), four per group, were intraperitoneally inoculated with 100 mL of K. pneumoniae strains at the mid-logarithmic growth phase (Mizuta et al., 1983). Before inoculation, K. pneumoniae strains were washed twice with normal saline and centrifuged at 10,621 × g for 4 min. A 0.6 McFarland standard equivalent to 2.0 × 10 8 colony forming units (CFU)/mL was prepared. The final inoculation was 10 2 -10 7 CFU/mL. The mice were observed for 14 d after inoculation. LD 50 was determined according to a previous study (Reed and Muench, 1938). Strains NTUH-K2044 and HS11286 were used as positive and negative controls. K. pneumoniae strains with LD 50 ≤ 10 times of that of NTUH-K2044 were regarded as hypervirulent; those with LD 50 > 10 times of that of NTUH-K2044 were denoted as hypovirulent.

Statistical Analysis
GraphPad Prism 8 software (GraphPad Software Inc., Sand Diego, CA, USA) was used to perform Chi-square test, oneway ANOVA, and Kruskal-Wallis test between groups; p < 0.05 was considered significant.
In total, 13 key virulence genes in K. pneumoniae were investigated in this study; Figure 2 shows a remarkable divergence in their distribution. The extremely high detection rates of fimH, mrkD, entB, and wzi indicate ubiquitous production of type 1 fimbriae, type 2 fimbriae, enterobactins, and capsules. Another siderophore gene irp2 was present in a large proportion of the strains (76.1%), and is typically carried by ICEKp1 in the chromosome; The remaining 2 siderophore genes iucA and iroN showed detection rates of approximately 20.0-30.0%, and are usually harboured by pK2044 and pLVPK-like virulence plasmids (Struve et al., 2015). The virulence genes peg-344, p-rmpA, and p-rmpA2 are also present in pK2044-and pLVPK-like virulence plasmids (Russo and Marr, 2019), and therefore yielded similar detection rates to those of iucA and iroN (p > 0.05). The c-rmpA gene is present in the chromosome and often found in PLA specimens (Hsu et al., 2011). No such specimens were included in this study, which may explain the detection rate of 0.0%. The low detection rate of wzy-K1 (6.4%) suggests the rarity of K1 K. pneumoniae in clinical practice.
Various molecular factors were evaluated to screen for CR-HvKP and Hv-CRKP. Zhang et al. previously evaluated iucA,iroN,rmpA,and rmpA2 to check for the presence of virulence plasmids, and performed the G. mellonella lethality test to identify CR-HvKP, which yielded a rate of 5.2% (55/1052) for CR-HvKP (Zhang et al., 2020). We previously defined HvKP as: positive wzy-K1, ≥3 positive siderophore genes (entB, irp2, iroN, and iucA), or ≥1 positive capsule-regulating genes (p-rmpA2, c-rmpA/A2, and p-rmpA), and estimated a rate of 5.6% (29/521) for Hv-bla KPC (+)-KP (Shon et al., 2013), which was also applied in this study. Harada et al. defined HvKP as strains carrying virulence genes, rmpA, rmpA2, iroBCDN, iucABCD, or iutA (Harada et al., 2019). Russo et al. confirmed that the G. mellonella lethality experiment cannot accurately differentiate HvKP from cKP (Russo and MacDonald, 2020). Thus, the mouse lethality test may represent the only approach to determine the exact prevalence of CR-HvKP/Hv-CRKP. In this study, only two bla KPC (+)-HvKP strains, but no Hv-bla KPC (+)-KP strains, were eventually confirmed using a mouse lethality test, showing a rate of 0.5% (2/436) for CR-HvKP which was far lower than that reported in other studies (Zhang et al., 2020;Hu et al., 2021). Owing to the predominance of KPC-induced carbapenem resistance , the actual prevalence of CR-HvKP should be approximately 0.5-1.0% in mainland China in 2017. The considerable difference in prevalence rates determined between this study and other reports highlights the need to elucidate why such biomarkers are not reliable and the difference between mouse and G. mellonella lethality tests. Zhang et al., reported that only one K. pneumoniae strain has been confirmed as CR-HvKP among three strains that harbour rmpA based on a mouse lethality test (Zhang et al., 2015). The fact that rmpA genes are nonfunctional in cKP may be attributed to different genetic backgrounds, although rmpA-related genes, such as kvrA, kvrB, and rcsB (Palacios et al., 2018;Walker et al., 2019), were found to be widely distributed in both cKP and HvKP (data not shown).
Although Hv-CRKP and CR-HvKP strains are currently emerging worldwide (Gu et al., 2018;Karlsson et al., 2019), our study revealed that the emergence of CR-HvKP is a relatively greater concern owing to its prevalence. In this study, two confirmed bla KPC (+)-HvKP strains were found, including JS184 (K2) and TZ20 (K20), which showed no fitness cost, no hypercapsule production, and high expression of galF which is responsible for the synthesis of capsule precursor (Peng et al., 2018;Walker et al., 2019). However, JS184 showed a positive string test and exopolysaccharide production in contrast to TZ20. The reason for this is not known. In addition, JS184 and TZ20 also showed different ST and irp2 expression.
This study had a few limitations. First, only typical siderophore genes were referred to, but not their expression. Second, capsule staining is not sufficient to differentiate capsules of various thicknesses, which may impact virulence.
Taken together, our findings indicate that CR-HvKP may emerge more often than Hv-CRKP; the former accounted for less than 1.0% of the strains evaluated via mouse lethality tests among clinical K. pneumoniae strains in mainland China in 2017.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material.

ETHICS STATEMENT
The animal study was reviewed and approved by the Institutional Animal Care and Use Committee of the School of Pharmacy, Fudan University (Shanghai, China).