Linden T. Hu

The mouse model of tick-mediated Borrelia burgdorferi sensu lato ( Bb sl) challenge is integral to the preclinical evaluation of Lyme disease vaccines. The standard measure of vaccine efficacy is the recovery of viable spirochetes from different mouse tissues, monitored weekly for a month post challenge. Serological tests (ELISA and Western blots) with B. burgdorferi lysates, recombinant proteins, or peptides are often used as alternative or confirmatory readouts. While there are several reports suggesting that culture positivity and serology are strongly correlated, a systematic study addressing this issue is lacking. In this report, we took advantage of available data sets from large cohorts of naive and OspA actively and passively immunized mice that had been challenged with Ixodes scapularis nymphs carrying Bb sl strain B31, euthanized 3 weeks later, then assessed for culture positivity across different tissue types (ear, skin, ankle, bladder, and heart) and seropositivity using an OspC and DbpB 2-plex microsphere immunoassay (MIA) and a vmp-like sequence C6 monoplex MIA. Our results demonstrate a 96.5%–97% agreement between culture positivity and seroconversion using the OspC/DbpB and C6 MIAs, suggesting that serology provides a near-perfect readout of disseminated spirochetal infection. While some instances of culture and serological discordance could be attributed to technical errors, others likely reflect aspects of infectivity/immunological dynamics, including the possibility of abortive (non-sterilizing) B. burgdorferi infections. We conclude that serological testing provides a highly accurate and practical measure of disseminated B. burgdorferi infection in mouse models, crucial for the efficient preclinical assessment of Lyme disease vaccines. IMPORTANCE: The laboratory mouse ( Mus musculus ) plays an indispensable role in the testing of candidate vaccines for Lyme disease, an emerging tick-borne infection caused by the spirochete, Borrelia burgdorferi . While laboratory mice do not develop Lyme disease per se, B. burgdorferi proliferates in skin tissues following a tick bite, then disseminates to other organs over the course of days to weeks. In this report, we examined the concordance between two widely used methodologies to assess disseminated B. burgdorferi infection and offer recommendations to standardize testing across laboratories.

Toll-like receptor 2/1 (TLR2/1) heterodimers detect triacylated lipoproteins. Although TLR2/1 complexes localize to both the plasma membrane and endosome, the functional contributions of compartment-specific signaling are not completely defined. A common human TLR1 polymorphism (rs5743618), at which one allele substitutes isoleucine for serine adjacent to the transmembrane domain, restricts TLR1 intracellularly and is associated with an increased risk of antibiotic-refractory Lyme arthritis. This mis-localization provides a model to examine how receptor positioning influences downstream signaling. Here, we investigated the rs5743618 polymorphism’s effects on TLR1 localization and compartment-specific NF-κB activation. Using a CD14+ TLR2+ HEK cell reporter system that enables transient expression of both allelic variants, we confirmed that the polymorphism is excluded from the cell surface and retained intracellularly, whereas the ancestral allele localizes to both compartments. Functionally, the ancestral allele signaled from both compartments, while the derived allele signaled exclusively from endosomes. Spatially restricted signaling exhibited influence on IkBα degradation and NF-κB activation compared with the ancestral allele. This activation was dependent on MyD88 localized at ligand-containing endosomes, establishing a role for this adapter in intracellular TLR2/1 signaling. These findings provide insight into how this common polymorphism impacts inflammatory outcomes.

Background: Electronic health records provide opportunities for improving Lyme disease surveillance by collecting detailed clinical information and timely, scalable disease incidence estimation. Medical record review can improve understanding of clinical manifestations, how likely and possible cases compare, and potential misclassification of possible cases. Methods: The SubLyme Network is a collaboration among 5 healthcare systems, a coordinating center, and the Centers for Disease Control and Prevention. The sites each conducted stratified random sampling of 500 potential acute Lyme disease episodes for medical record review from patients with a diagnosis, test order, and/or appropriate antibiotic in 2022 or 2023. Reviewers abstracted data, including from free text notes, to classify patients as likely or possible Lyme disease and by clinical manifestation. Multivariable logistic regression was used to identify differences between possible compared to likely Lyme disease cases. Results: Among 2500 reviews, 880 patients were categorized as having likely or possible new onset, active Lyme disease. The patient proportion identified as having a specific clinical manifestation ranged from 61% to 76% across sites. Of likely and possible cases, 55.9%, 7.8%, 1.3%, and 3.1% were diagnosed with erythema migrans, Lyme arthritis, Lyme carditis, and neuroborreliosis, respectively. Possible cases differed from likely cases in having more nonspecific symptoms, non-Lyme disease diagnoses, provider uncertainty, and undocumented provider decision-making. Conclusions: Medical record review is expected to be more accurate than computable phenotypes for Lyme disease identification and staging, yet likely and possible cases were challenging to distinguish, and a sizable proportion could not have clinical manifestation assigned. Findings provide direction for computable phenotype refinement to improve case identification. KEY WORDS: clinical manifestations, electronic health records, epidemiology, Lyme disease, surveillance

Background: Lyme disease is the most common vector-borne illness in the United States. The limitations of traditional surveillance strategies for Lyme disease affect the ability to reliably track its burden and evaluate interventions. The US Centers for Disease Control and Prevention (CDC) established the Surveillance Based Lyme Disease Network (SubLyme) in September 2023 to strengthen Lyme disease surveillance and research using electronic health record (EHR) data. Objective: SubLyme has three primary objectives: (1) to establish and evaluate criteria for identifying Lyme disease cases in EHR data (ie, create computable phenotypes [CPs]) that can be scaled across diverse health systems, (2) to estimate Lyme disease incidence, and (3) to describe Lyme disease incidence by key demographics. Secondary objectives are to develop CPs that distinguish between acute and disseminated Lyme disease, identify clinical manifestations, and support future research efforts. This paper describes SubLyme, its structure, and its methods. Methods: SubLyme includes 5 health systems in 3 US regions with a high risk of Lyme disease: Geisinger, in Pennsylvania; Marshfield Clinic Health System, in Wisconsin; and Mass General Brigham, Tufts Medical Center, and MaineHealth in New England. The network is administered by a coordinating center (Westat) and the US CDC. SubLyme is evaluating the validity of EHR-based CP definitions for Lyme disease. CP performance is assessed by measuring sensitivity, specificity, positive predictive value, and negative predictive value against manually abstracted medical charts. Each site identified a cohort of patients with any Lyme disease element in their EHR (Lyme disease diagnosis code, Lyme disease laboratory test order, and Lyme-appropriate antibiotic order) during 2022 to 2023 and selected 500 charts for manual review as the gold standard against which CP performance was evaluated. SubLyme will use the Lyme disease CPs to generate incidence rates for Lyme disease overall and for various subgroups. Results: SubLyme identified 332,256 patients with at least 1 Lyme disease element in their record from more than 4.6 million patients. Of these patients, 55.6% (n=184,734) were female, 87.9% (n=292,053) were White, and 90.8% (n=301,688) were non-Hispanic. More than half of the patients only had a Lyme-appropriate medication order (n=177,425, 53.4%) and 35.8% (n=118,948) only had a Lyme disease test order. The most common combination was a medication order with a laboratory test order (n=22,926, 6.9%), followed by a combination of a diagnosis, test, and medication order (n=5316, 1.6%). Conclusions: SubLyme is well positioned to advance Lyme disease surveillance using EHR data across multiple health systems. The exploration of new surveillance methods in Lyme disease is critical as disease frequency increases and the geography expands. An EHR-based approach to surveillance has the potential to overcome challenges of current surveillance strategies and to accelerate Lyme disease research. International Registered Report Identifier (IRRID): DERR1-10.2196/94921

Machine learning (ML) has accelerated molecular discovery, yet training models to generalize to out-of-distribution (OOD) chemical spaces remains fundamentally constrained by the high cost of experimental validation. In antibiotic discovery, where whole-cell phenotypic high throughput screening (HTS) is resource-intensive, iterative ML-guided compound selection, or Active Learning (AL), offers a pathway to efficiently navigate available chemical spaces. However, the algorithmic tradeoffs between prioritizing compound novelty (exploration), predicted bioactivity (exploitation), and their impact on OOD generalizability remain unresolved for noisy, whole-cell biological systems. In this work, we systematically evaluate three AL strategies for whole-cell bacterial bioactivity and benchmark their effects on model accuracy, hit rate, and OOD performance. Using retrospective simulations on Mycobacterium tuberculosis HTS data, we identify an optimal AL strategy that balances predicted hit/non-hit novelty with overall hit rate. We then integrate the strategy in a closed-loop Borrelia burgdorferi antibiotic discovery HTS campaign. The AL-guided approach successfully increased the experimental screening hit rate five-fold (from a 0.2% rate within investigator-selected plates to 1.0%). Further, when the trained model was applied in prospective in silico selection of highly diverse compounds across multiple bacterial species, the AL-trained whole-cell inhibition predictor demonstrates 53-fold enrichment over investigator-directed screening (11.0% experimental validation of predicted hits). Of these, 100% demonstrated the intended narrow spectrum activity for Borrelia burgdorferi. These results demonstrate that calibrated AL strategies can overcome data acquisition bottlenecks and train generalizable property predictors able to extrapolate to OOD molecules.

Acute Lyme disease is caused by a Borrelia infection and typically responds to antibiotic treatment, but post-infectious chronic symptoms are common. The precise origin of these post-treatment symptoms is unknown; dysregulation of immune responses raised during the initial infection is likely to contribute. Antilipid antibodies are associated with several autoimmune diseases and have been shown to arise in both acute and post-treatment Lyme disease. To assess the potential contribution of antilipid antibodies to Lyme disease pathology and their possible use as biomarkers of both acute and chronic disease, we performed a survey of patient sera for antilipid antibodies during and after Borrelia burgdorferi infection. Results were similar in cross-sectional and longitudinal samples drawn from two different biobanks. Three antiphospholipid antibodies were elevated during infection, with two (antiphosphatidic acid and antiphosphatidylserine) significantly elevated even at the day of diagnosis before seroconversion on conventional tests. A subset of patients with chronic symptoms is identifiable by persistent elevation in antiphosphatidylserine; these antibodies found in post-treatment Lyme disease were not found in a panel of sera from patients with look-alike autoimmune disorders. Persistent elevation in antiphosphatidylserine may drive autoimmune-like symptoms of Lyme disease in some patients and could serve as a biomarker for chronic disease. The detection of antiphospholipid antibodies induced early in infection could also improve the diagnosis of acute infections.

Sam R Telford, III, Gary P Wormser, Kenneth Dardick, Luis A Marcos, Rafal Tokarz, Linden T Hu, Adriana R Marques; Vectors are not just culture vessels, Cli

BACKGROUND Lyme disease is the most common vector-borne illness in the United States. The limitations of traditional surveillance strategies for Lyme disease affect the ability to reliably track its burden and evaluate interventions. The US Centers for Disease Control and Prevention (CDC) established the Surveillance Based Lyme Disease Network (SubLyme) in September 2023 to strengthen Lyme disease surveillance and research using electronic health record (EHR) data. OBJECTIVE SubLyme has three primary objectives: (1) to establish and evaluate criteria for identifying Lyme disease cases in EHR data (ie, create computable phenotypes [CPs]) that can be scaled across diverse health systems, (2) to estimate Lyme disease incidence, and (3) to describe Lyme disease incidence by key demographics. Secondary objectives are to develop CPs that distinguish between acute and disseminated Lyme disease, identify clinical manifestations, and support future research efforts. This paper describes SubLyme, its structure, and its methods. METHODS SubLyme includes 5 health systems in 3 US regions with a high risk of Lyme disease: Geisinger, in Pennsylvania; Marshfield Clinic Health System, in Wisconsin; and Mass General Brigham, Tufts Medical Center, and MaineHealth in New England. The network is administered by a coordinating center (Westat) and the US CDC. SubLyme is evaluating the validity of EHR-based CP definitions for Lyme disease. CP performance is assessed by measuring sensitivity, specificity, positive predictive value, and negative predictive value against manually abstracted medical charts. Each site identified a cohort of patients with any Lyme disease element in their EHR (Lyme disease diagnosis code, Lyme disease laboratory test order, and Lyme-appropriate antibiotic order) during 2022 to 2023 and selected 500 charts for manual review as the gold standard against which CP performance was evaluated. SubLyme will use the Lyme disease CPs to generate incidence rates for Lyme disease overall and for various subgroups. RESULTS SubLyme identified 332,256 patients with at least 1 Lyme disease element in their record from more than 4.6 million patients. Of these patients, 55.6% (n=184,734) were female, 87.9% (n=292,053) were White, and 90.8% (n=301,688) were non-Hispanic. More than half of the patients only had a Lyme-appropriate medication order (n=177,425, 53.4%) and 35.8% (n=118,948) only had a Lyme disease test order. The most common combination was a medication order with a laboratory test order (n=22,926, 6.9%), followed by a combination of a diagnosis, test, and medication order (n=5316, 1.6%). CONCLUSIONS SubLyme is well positioned to advance Lyme disease surveillance using EHR data across multiple health systems. The exploration of new surveillance methods in Lyme disease is critical as disease frequency increases and the geography expands. An EHR-based approach to surveillance has the potential to overcome challenges of current surveillance strategies and to accelerate Lyme disease research. INTERNATIONAL REGISTERED REPORT DERR1-10.2196/94921

Background: Some patients report nonspecific symptoms after antibiotic therapy for Lyme disease (LD), raising questions about ongoing infection, despite no compelling evidence. We investigated whether xenodiagnosis could detect Borrelia burgdorferi in such patients and if positive results correlated with symptoms. Methods: Participants were adults who completed antibiotic treatment for LD 3–12 months earlier (posttherapy, n = 40) or had persistent symptoms for ≥12 months after treatment, (posttreatment LD symptoms [PTLDS], n = 20). Controls included 1 patient with erythema migrans (EM), 1 patient with untreated Lyme arthritis (LA), and 9 healthy volunteers (HV). Participants had 25–30 larval Ixodes scapularis ticks placed; ticks were collected 3–6 days later and tested for B. burgdorferi. The primary analysis evaluated if B. burgdorferi detection by xenodiagnosis was associated with persistence of symptoms in patients during the first year after treatment. This trial is registered with ClinicalTrials.gov, NCT02446626. Results: Recovered ticks included 402 from posttherapy, 314 from PTLDS, 30 from the EM patient, 11 from the LA patient, and 80 from HV. All ticks tested negative for B. burgdorferi except for 1 tick from a recovered patient. An unplanned interim analysis led to the early termination of the study for futility. Conclusions: Xenodiagnosis with larval I. scapularis ticks showed no evidence of B. burgdorferi in most patients after treatment, irrespective of symptoms. This may be due to the absence of bacteria or to the low sensitivity of the technique in humans. This method is unlikely to detect persistent B. burgdorferi infection in humans, and further research on the use of xenodiagnosis is unwarranted.

Lyme disease spirochetes are maintained in natural reservoirs before spilling over into human populations. Targeting these reservoirs with vaccinations or antibiotics could impact the Borrelia burgdorferi enzootic cycle and reduce the risk of human Lyme disease. In this work we report that the narrow-spectrum antibiotic hygromycin A is sufficient to disrupt B burgdorferi transmission from the main eastern US reservoir, Peromyscus leucopus, to ticks. Additionally, hygromycin A–containing baits can clear B burgdorferi from P leucopus. These studies lay the foundation for the use of hygromycin A as a reservoir-targeted antibiotic to eradicate B burgdorferi in the wild.