Hypervirulent Klebsiella pneumoniae: Mechanisms, Genetics, and Disease

Hypervirulent Klebsiella pneumoniae, often abbreviated hvKp, is a pathotype of a common gut bacterium that can cause devastating, fast-moving infections in people who are otherwise healthy. Unlike the classical form of Klebsiella pneumoniae that mostly threatens hospitalized or immunocompromised patients, hvKp was first recognized in the 1980s as the leading cause of a distinctive syndrome: community-acquired pyogenic liver abscess, often with spread to the eyes, lungs, or brain.¹ What makes this organism particularly alarming is a genetic toolkit that lets it scavenge iron from the host, evade immune defenses, and seed infections in distant organs, sometimes all at once.

What Sets Hypervirulent Strains Apart

Standard Klebsiella pneumoniae is an opportunist. It usually causes trouble in hospitals: urinary tract infections related to catheters, ventilator-associated pneumonia, bloodstream infections after surgery. These infections overwhelmingly occur in people whose defenses are already compromised. Hypervirulent strains break that pattern. They infect people in the community, often without any obvious predisposing condition other than diabetes, and they produce infections that are far more aggressive.

The hallmark presentation is a monomicrobial liver abscess, meaning the abscess contains only Klebsiella and develops without underlying biliary disease like gallstones. On imaging, these abscesses tend to be single, confined to one lobe, and solid-looking or multiloculated. Blood clots in the portal vein occur in up to a third of cases.² But the feature that makes hvKp most feared is its tendency to spread from the liver to distant sites while the original infection is still being treated.

How Infections Spread Through the Body

The current understanding is that hvKp likely enters the liver through the portal vein after colonizing the intestine. Patients typically show no obvious skin or mucosal lesions that would explain how the bacteria got in.³ Once in the bloodstream, the bacteria can disseminate to the eyes, lungs, central nervous system, and other organs.

Endogenous endophthalmitis, an infection inside the eye seeded by bacteria traveling through the blood, was one of the first recognized complications and remains one of the most devastating. It occurs in roughly 5% of patients with hvKp bacteremia, and the outcomes are grim: visual symptoms progress rapidly even with aggressive antibiotic treatment, and vision drops to light perception or worse in about 89% of affected patients.² Meningitis, pneumonia, and abscesses in non-hepatic sites round out the spectrum. The speed of dissemination means delayed diagnosis at any of these sites dramatically worsens the prognosis.³

The Iron-Stealing Strategy at the Heart of Virulence

Iron is essential for bacterial growth, and the human body keeps free iron at vanishingly low levels precisely to starve pathogens. Bacteria get around this by producing siderophores, small molecules that bind iron tightly and ferry it back to the cell. Klebsiella pneumoniae can produce up to four different siderophores, but in hypervirulent strains one of them dominates: aerobactin.

Experimental work has shown that aerobactin accounts for the overwhelming majority of the extra siderophore production that distinguishes hvKp from classical strains, and this isn’t simply because the gene is present in more copies.⁴ When researchers knocked out the other three siderophore systems (enterobactin, salmochelin, and yersiniabactin) individually or in combination, the bacteria grew and survived just as well in human serum and in animal infection models. Removing aerobactin, on the other hand, crippled the bacteria’s ability to thrive in those same environments.⁵ This makes aerobactin the single most critical siderophore for hvKp virulence.

The Genetic Toolkit

The genes responsible for hypervirulence don’t live on the main bacterial chromosome alone. Many sit on a large virulence plasmid, often a roughly 180-kilobase structure related to the well-studied pLVPK plasmid. This plasmid carries the genes for aerobactin production (the iuc locus), genes that regulate the thick mucoid capsule (rmpA and rmpA2), and additional iron-uptake systems.⁶ Because these genes are plasmid-borne, they can potentially transfer between bacterial strains, a point with serious implications for resistance.

Capsule type also matters. The K1 and K2 capsular serotypes are the most closely associated with hypervirulence. In studies comparing hypermucoviscous strains to non-hypermucoviscous strains, K1 and K2 were significantly more common in the hypermucoviscous group, alongside higher rates of the aerobactin, rmpA, and rmpA2 genes.⁷ Research on the outer-membrane protein Wzi has shown it helps the bacterium resist killing by human serum in ways that go beyond just anchoring the capsule in place, suggesting the immune-evasion machinery has multiple layers.⁸

The chromosomal integrative element ICEKp10, which encodes both yersiniabactin and colibactin, adds another dimension. Colibactin is a genotoxin that damages DNA, and it appears to play a key role not in systemic virulence but in gut colonization, helping the bacterium outcompete other microbes in the intestine.⁹ A second toxin encoded by the same element, microcin E492, amplifies colibactin’s colonization advantage.

The Gut as a Silent Reservoir

Before hvKp causes a liver abscess or bloodstream infection, it has to be living somewhere in the body. Mounting evidence points to the gastrointestinal tract. Surveys of healthy adults in China have found that roughly 4% to 5% carry hvKp in their stool, and phylogenetic analysis shows these gut strains are closely related to the ones recovered from fatal clinical infections.¹⁰ The implication is that carriage in the gut precedes disease, and some event — perhaps a shift in gut ecology, immune suppression, or mucosal injury — allows the bacteria to translocate into the portal circulation.

The colibactin-driven colonization mechanism described above helps explain how hvKp can persist in the gut against competition from resident microbes. Strains lacking colibactin lost their foothold in the intestine over time in animal models, while strains with an intact colibactin gene maintained stable colonization.⁹

Why Diabetes Tips the Scales

Diabetes, especially type 2, shows up repeatedly as a risk factor for hvKp liver abscess. The relationship goes deeper than general immune compromise. When hvKp is grown in high-glucose conditions mimicking poorly controlled diabetes, the bacteria ramp up expression of the rmpA gene (which thickens the capsule) and the ompA gene, and they become better at surviving in human serum.¹¹ High blood sugar, in other words, directly makes the bacteria more dangerous.

On the host side, neutrophils from people with type 2 diabetes show a specific defect in killing carbapenem-resistant hvKp. Paradoxically, these neutrophils actually produce more neutrophil extracellular traps (web-like structures that snare and kill bacteria) than healthy controls. The problem isn’t forming the traps; it’s that the traps inflict less surface damage on the bacteria once they are caught.¹² Immune cells from diabetic patients also show mildly reduced production of certain cytokines that coordinate the broader immune response against hvKp, and a slightly diminished ability to control bacterial replication inside cells.¹³ The combination of bacteria that become more virulent and a host whose immune response is weakened at multiple points creates a particularly dangerous situation.

When Hypervirulence Meets Antibiotic Resistance

For decades, one piece of somewhat reassuring news about hvKp was that most strains remained susceptible to common antibiotics, including carbapenems, the drugs of last resort for serious Gram-negative infections. Antibiotic resistance and hypervirulence traveled in different genetic lineages: multidrug-resistant clones like ST258, ST147, and ST101 predominated in Western healthcare settings, while hypervirulent clones like ST23, ST65, and ST86 predominated in Asia.¹⁴

That separation is collapsing. Convergent strains, those carrying both virulence plasmids and carbapenem-resistance genes, are being reported with increasing frequency. In one well-studied cluster in China, ST11 strains (a lineage classically associated with resistance, not virulence) were found carrying both a carbapenemase-producing resistance plasmid and an rmpA2-bearing virulence plasmid.¹⁵ In Singapore, hvKp strains with K1, K2, and K20 capsular serotypes acquired a plasmid carrying the carbapenemase gene blaKPC-2.¹⁶ Convergent strains have also caused hospital outbreaks, including cases of extensively drug-resistant hvKp in intensive care units.¹⁷

The earliest documented case of carbapenem-resistant hvKp in the dominant hypervirulent lineage ST23 dates to 2013, when a strain carrying blaKPC-2 alongside the core virulence genes rmpA, rmpA2, iroN, iucA, and pagO was isolated in Zhejiang province, China.¹⁸ Global surveillance data confirm that ST23, ST65, and ST86 remain the lineages with the highest proportion of hypervirulent isolates, and the overlap with multidrug resistance is growing, particularly in Asia.¹⁹

How Hospital Outbreaks Unfold

When convergent strains enter healthcare facilities, the results can be severe. In one documented outbreak at a large teaching hospital, five patients in overlapping intensive care stays developed infections with colistin- and carbapenem-resistant hvKp. The chain of transmission was traced through shared ICU time, shared consulting physicians, and shared medical teams. Patients ranged from their twenties to their late sixties and included people recovering from neurosurgery, bone marrow transplantation, and severe pneumonia. Isolates were recovered from sputum and urine after extended stays involving mechanical ventilation or indwelling catheters.²⁰ This kind of outbreak is particularly difficult to manage because the usual last-line antibiotics are ineffective, and the organism’s virulence traits make infection more likely to progress.

Identifying Hypervirulence in the Lab

One longstanding clinical test for hypervirulence is the string test: if a viscous string stretches more than five centimeters when a colony is touched with a loop, the strain is considered hypermucoviscous, a feature historically linked to hvKp. But “hypermucoviscous” and “hypervirulent” are not the same thing. The string test captures the thick capsule phenotype, but some genuinely hypervirulent strains are not hypermucoviscous, and some hypermucoviscous strains lack critical virulence genes.

Genetic biomarkers perform better. A study evaluating several candidate markers found that the genes iucA (aerobactin synthesis), iroB (salmochelin synthesis), rmpA, rmpA2, and a plasmid gene called peg-344 all achieved greater than 95% diagnostic accuracy for identifying hvKp strains. The string test, by comparison, reached only about 90% accuracy.²¹ The practical problem is that most clinical microbiology labs do not routinely test for these genes. Clinicians often have to rely on the clinical picture — a community-acquired monomicrobial liver abscess in a patient with or without diabetes, especially with metastatic complications — to suspect hvKp while awaiting confirmatory results.

Treatments Beyond Conventional Antibiotics

For strains that remain antibiotic-susceptible, standard therapy works. The crisis is what to do when resistance catches up with virulence. Several alternative approaches are under active investigation.

Bacteriophage therapy, which uses viruses that specifically kill bacteria, has shown promise in preclinical work. Phages and phage cocktails have demonstrated activity against hvKp strains including biofilm-forming and carbapenem-resistant isolates. Antimicrobial peptides derived from various biological sources, including scorpion venom, have also shown efficacy in animal models.²²

Monoclonal antibodies targeting the Klebsiella capsule represent another avenue. In mouse experiments, two antibodies protected against both sepsis and lung infection caused by hvKp. In mice already colonized with hvKp in the gut, antibody treatment reduced dissemination of bacteria from the intestine to lymph nodes and organs. Imaging showed the antibodies promoted efficient engulfment and clearance of bacteria by immune cells in the liver.²³ None of these approaches have reached routine clinical use, but they illustrate the seriousness with which the field is treating the prospect of untreatable hvKp infections.

Animal Reservoirs and the One Health Dimension

The story of hvKp extends beyond human medicine. Surveys of diarrheic farm animals have detected both hvKp and carbapenem-resistant hvKp across multiple species. In one study, the overall rate of hvKp among Klebsiella pneumoniae isolates from sick farm animals was about 8%, and carbapenem-resistant hypervirulent strains were found in all animal species examined.²⁴ These findings raise the possibility that farm animals serve as a reservoir where virulence and resistance genes can mingle before jumping back to humans, though the frequency and direction of actual cross-species transmission remain poorly characterized.

Even the evolutionary origins of hvKp’s dominant lineage hint at an animal connection. Genomic analysis of the CG23 clonal group, the most important liver-abscess-causing lineage, estimated that the overall population dates back to around 1878, while its main liver-abscess sublineage CG23-I emerged around 1928 after acquiring ICEKp10. Nested within CG23-I is an equine sublineage, estimated to have diverged around 1972.²⁵ The presence of horse-adapted strains within a human-disease lineage underscores that hvKp moves across host species, a reality that complicates surveillance efforts and any future vaccination strategy.

Research Models and Their Limits

Studying hvKp requires animal models that reliably distinguish it from classical strains. The standard laboratory mouse model does this well: when infected with defined panels of hvKp and classical strains, outbred mice show clearly different mortality patterns. But cheaper and faster models do not always work as substitutes. The wax moth larva, widely used as a quick screen for bacterial virulence, failed to differentiate between hypervirulent and classical Klebsiella strains in head-to-head comparisons.²⁶ This matters because researchers screening large collections of clinical isolates for hypervirulence need a model that matches the clinical reality, and using the wrong one could lead to misclassification.