Generation of nanobodies from transgenic 'LamaMice' lacking an endogenous immunoglobulin repertoire.

Due to their exceptional solubility and stability, nanobodies have emerged as powerful building blocks for research tools and therapeutics. However, their generation in llamas is cumbersome and costly. Here, by inserting an engineered llama immunoglobulin heavy chain (IgH) locus into IgH-deficient mice, we generate a transgenic mouse line, which we refer to as 'LamaMouse'. We demonstrate that LamaMice solely express llama IgH molecules without association to Igκ or λ light chains. Immunization of LamaMice with AAV8, the receptor-binding domain of the SARS-CoV-2 spike protein, IgE, IgG2c, and CLEC9A enabled us to readily select respective target-specific nanobodies using classical hybridoma and phage display technologies, single B cell screening, and direct cloning of the nanobody-repertoire into a mammalian expression vector. Our work shows that the LamaMouse represents a flexible and broadly applicable platform for a facilitated selection of target-specific nanobodies.

Structural and functional characterization of nanobodies that neutralize Omicron variants of SARS-CoV-2.

The Omicron strains of SARS-CoV-2 pose a significant challenge to the development of effective antibody-based treatments as immune evasion has compromised most available immune therapeutics. Therefore, in the 'arms race' with the virus, there is a continuing need to identify new biologics for the prevention or treatment of SARS-CoV-2 infections. Here, we report the isolation of nanobodies that bind to the Omicron BA.1 spike protein by screening nanobody phage display libraries previously generated from llamas immunized with either the Wuhan or Beta spike proteins. The structure and binding properties of three of these nanobodies (A8, H6 and B5-5) have been characterized in detail providing insight into their binding epitopes on the Omicron spike protein. Trimeric versions of H6 and B5-5 neutralized the SARS-CoV-2 variant of concern BA.5 both in vitro and in the hamster model of COVID-19 following nasal administration. Thus, either alone or in combination could serve as starting points for the development of new anti-viral immunotherapeutics.

Single domain antibodies from camelids in the treatment of microbial infections.

Infectious diseases continue to pose significant global health challenges. In addition to the enduring burdens of ailments like malaria and HIV, the emergence of nosocomial outbreaks driven by antibiotic-resistant pathogens underscores the ongoing threats. Furthermore, recent infectious disease crises, exemplified by the Ebola and SARS-CoV-2 outbreaks, have intensified the pursuit of more effective and efficient diagnostic and therapeutic solutions. Among the promising options, antibodies have garnered significant attention due to their favorable structural characteristics and versatile applications. Notably, nanobodies (Nbs), the smallest functional single-domain antibodies of heavy-chain only antibodies produced by camelids, exhibit remarkable capabilities in stable antigen binding. They offer unique advantages such as ease of expression and modification and enhanced stability, as well as improved hydrophilicity compared to conventional antibody fragments (antigen-binding fragments (Fab) or single-chain variable fragments (scFv)) that can aggregate due to their low solubility. Nanobodies directly target antigen epitopes or can be engineered into multivalent Nbs and Nb-fusion proteins, expanding their therapeutic potential. This review is dedicated to charting the progress in Nb research, particularly those derived from camelids, and highlighting their diverse applications in treating infectious diseases, spanning both human and animal contexts.

Nemabiome metabarcoding shows a high prevalence of Haemonchus contortus and predominance of Camelostrongylus mentulatus in alpaca herds in the northern UK.

Gastrointestinal nematodes (GINs) are a common threat faced by pastoral livestock. Since their major introduction to the UK in the early 1990s, South American camelids have been cograzed with sheep, horses, and other livestock, allowing exposure to a range of GIN species. However, there have been no molecular-based studies to investigate the GIN populations present in these camelids. In the current study, we sampled nine alpaca herds from northern England and southern Scotland and used high-throughput metabarcoded sequencing to describe their GIN species composition. A total of 71 amplicon sequence variants (ASVs) were identified representing eight known GIN species. Haemonchus contortus was the most prevalent species found in almost all herds in significant proportions. The identification of H. contortus in other livestock species is unusual in the northern UK, implying that alpacas may be suitable hosts and potential reservoirs for infection in other hosts. In addition, the camelid-adapted GIN species Camelostrongylus mentulatus was identified predominantly in herds with higher faecal egg counts. These findings highlight the value of applying advanced molecular methods, such as nemabiome metabarcoding to describe the dynamics of gastrointestinal nematode infections in novel situations. The results provide a strong base for further studies involving cograzing animals to confirm the potential role of alpacas in transmitting GIN species between hosts.

Clostridium tanneri sp. nov., isolated from the faecal material of an alpaca.

A strictly anaerobic, Gram-stain-negative rod-shaped bacterium, designated A1-XYC3T, was isolated from the faeces of an alpaca (Lama pacos). On the basis of the results of a comparative 16S rRNA gene sequence analysis, the isolate was assigned to the genus Clostridium with the highest sequence similarities to Clostridium magnum DSM 2767T (96.8 %), Clostridium carboxidivorans P7T (96.3 %) and Clostridium aciditolerans JW/YJL-B3T (96.1 %). The average nucleotide identity between A1-XYC3T, C. magnum, C. carboxidivorans and C. aciditolerans was 77.4, 76.1 and 76.6  %, respectively. The predominant components of the cellular fatty acids of A1-XYC3T were C14 : 0, C16 : 0 and summed feature 10, containing C18:0/C17:0 cyclo. The DNA G+C content was 32.4 mol%. On the basis of biochemical, phylogenetic, genotypic and chemotaxonomic criteria, this isolate represents a novel species within Clostridium sensu stricto for which the name Clostridium tanneri sp. nov. is proposed. The type strain of this species is strain A1-XYC3T (=CCM 9376T=NRRL B-65691T).

History of Diversification and Adaptation from North to South Revealed by Genomic Data: Guanacos from the Desert to Sub-Antarctica.

The increased availability of quality genomic data has greatly improved the scope and resolution of our understanding of the recent evolutionary history of wild species adapted to extreme environments and their susceptibility to anthropogenic impacts. The guanaco (Lama guanicoe), the largest wild ungulate in South America, is a good example. The guanaco is well adapted to a wide range of habitats, including the Sechura Desert, the high Andes Mountains to the north, and the extreme temperatures and conditions of Navarino Island to the south. Guanacos also have a long history of overexploitation by humans. To assess the evolutionary impact of these challenging habitats on the genomic diversity, we analyzed 38 genomes (∼10 to 16×) throughout their extensive latitudinal distribution from the Sechura and Atacama Desert to southward into Tierra del Fuego Island. These included analyses of patterns of unique differentiation in the north and geographic region further south with admixture among L. g. cacsilensis and L. g. guanicoe. Our findings provide new insights on the divergence of the subspecies ∼800,000 yr BP and document two divergent demographic trajectories and to the initial expansion of guanaco into the more southern portions of the Atacama Desert. Patagonian guanacos have experienced contemporary reductions in effective population sizes, likely the consequence of anthropogenic impacts. The lowest levels of genetic diversity corresponded to their northern and western limits of distribution and some varying degrees of genetic differentiation. Adaptive genomic diversity was strongly linked with environmental variables and was linked with colonization toward the south followed by adaptation.

Development, High-Throughput Profiling, and Biopanning of a Large Phage Display Single-Domain Antibody Library.

Immunoglobulin G-based monoclonal antibodies (mAbs) have been effective in treating various diseases, but their large molecular size can limit their penetration of tissue and efficacy in multifactorial diseases, necessitating the exploration of alternative forms. In this study, we constructed a phage display library comprising single-domain antibodies (sdAbs; or "VHHs"), known for their small size and remarkable stability, using a total of 1.6 × 109 lymphocytes collected from 20 different alpacas, resulting in approximately 7.16 × 1010 colonies. To assess the quality of the constructed library, next-generation sequencing-based high-throughput profiling was performed, analyzing approximately 5.65 × 106 full-length VHH sequences, revealing 92% uniqueness and confirming the library's diverse composition. Systematic characterization of the library revealed multiple sdAbs with high affinity for three therapeutically relevant antigens. In conclusion, our alpaca sdAb phage display library provides a versatile resource for diagnostics and therapeutics. Furthermore, the library's vast natural VHH antibody repertoire offers insights for generating humanized synthetic sdAb libraries, further advancing sdAb-based therapeutics.

Perception of optical illusions in ungulates: insights from goats, sheep, guanacos and llamas.

Optical illusions have long been used in behavioural studies to investigate the perceptual mechanisms underlying vision in animals. So far, three studies have focused on ungulates, providing evidence that they may be susceptible to some optical illusions, in a way similar to humans. Here, we used two food-choice tasks to study susceptibility to the Müller-Lyer and Delboeuf illusions in 17 captive individuals belonging to four ungulate species (Lama guanicoe, Lama glama, Ovis aries, Capra hircus). At the group level, there was a significant preference for the longer/larger food over the shorter/smaller one in control trials. Additionally, the whole group significantly preferred the food stick between two inward arrowheads over an identical one between two outward arrowheads in experimental trials of the Müller-Lyer task, and also preferred the food on the smaller circle over an identical one on the larger circle in the experimental trials of the Delboeuf task. Group-level analyses further showed no significant differences across species, although at the individual level we found significant variation in performance. Our findings suggest that, in line with our predictions, ungulates are overall susceptible to the Müller-Lyer and the Delboeuf illusions, and indicate that the perceptual mechanisms underlying size estimation in artiodactyls might be similar to those of other species, including humans.

Functional insights of Tyr37 in framework region 2 directly contributing to the binding affinities and dissociation kinetics in single-domain VHH antibodies.

Single-domain VHH antibody is regarded as one of the promising antibody classes for therapeutic and diagnostic applications. VHH antibodies have amino acids in framework region 2 that are distinct from those in conventional antibodies, such as the Val37Phe/Tyr (V37F/Y) substitution. Correlations between the residue type at position 37 and the conformation of the CDR3 in VHH antigen recognition have been previously reported. However, few studies focused on the meaning of harboring two residue types in position 37 of VHH antibodies, and the concrete roles of Y37 have been little to be elucidated. Here, we investigated the functional states of position 37 in co-crystal structures and performed analyses of three model antibodies with either F or Y at position 37. Our analysis indicates that Y at position 37 enhances the dissociation rate, which is highly correlated with drug efficacy. Our findings help to explain the molecular mechanisms that distinguish VHH antibodies from conventional antibodies.

Development of nanobodies targeting hepatocellular carcinoma and application of nanobody-based CAR-T technology.

BACKGROUND: Chimeric antigen receptor T (CAR-T) cell therapy, as an emerging anti-tumor treatment, has garnered extensive attention in the study of targeted therapy of multiple tumor-associated antigens in hepatocellular carcinoma (HCC). However, the suppressive microenvironment and individual heterogeneity results in downregulation of these antigens in certain patients' cancer cells. Therefore, optimizing CAR-T cell therapy for HCC is imperative. METHODS: In this study, we administered FGFR4-ferritin (FGFR4-HPF) nanoparticles to the alpaca and constructed a phage library of nanobodies (Nbs) derived from alpaca, following which we screened for Nbs targeting FGFR4. Then, we conducted the functional validation of Nbs. Furthermore, we developed Nb-derived CAR-T cells and evaluated their anti-tumor ability against HCC through in vitro and in vivo validation. RESULTS: Our findings demonstrated that we successfully obtained high specificity and high affinity Nbs targeting FGFR4 after screening. And the specificity of Nbs targeting FGFR4 was markedly superior to their binding to other members of the FGFR family proteins. Furthermore, the Nb-derived CAR-T cells, targeting FGFR4, exhibited significantly enhanced anti-tumor efficacy in both experiments when in vitro and in vivo. CONCLUSIONS: In summary, the results of this study suggest that the CAR-T cells derived from high specificity and high affinity Nbs, targeting FGFR4, exhibited significantly enhanced anti-tumor efficacy in vitro and in vivo. This is an exploration of FGFR4 in the field of Nb-derived CAR-T cell therapy for HCC, holding promise for enhancing safety and effectiveness in the clinical treatment of HCC in the future.

Intra-horn insemination in the alpaca Vicugna pacos: Copulatory wounding and deep sperm deposition.

Alpacas (Vicugna pacos) are reported to be the rare mammal in which the penis enters the uterus in mating. To date, however, only circumstantial evidence supports this assertion. Using female alpacas culled for meat, we determined that the alpaca penis penetrates to the very tips of the uterine horns, abrading the tract and breaking fine blood vessels. All female alpacas sacrificed one hour or 24 hours after mating showed conspicuous bleeding in the epithelium of some region of their reproductive tract, including the hymen, cervix and the tips of each uterine horn, but typically not in the vagina. Unmated females showed no evidence of conspicuous bleeding. Histological examination of mated females revealed widespread abrasion of the cervical and endometrial epithelium, injuries absent in unmated females. Within one hour of mating, sperm were already present in the oviduct. The male alpaca's cartilaginous penis tip with a hardened urethral process is likely responsible for the copulatory abrasion. The entire female reproductive tract interacts with the penis, functioning like a vagina. Alpacas are induced ovulators, and wounding may hasten delivery of the seminal ovulation-inducing factor beta-NGF into the female's blood stream. There is no evidence of sexual conflict in copulation in alpaca, and thus wounding may also be one of a variety of mechanisms devised by mammals to induce a beneficial, short-term inflammatory response that stimulates blastocyst implantation, the uterine remodeling associated with placental development, and thus the success of early pregnancy.

Quality of Answers of Generative Large Language Models Versus Peer Users for Interpreting Laboratory Test Results for Lay Patients: Evaluation Study.

BACKGROUND: Although patients have easy access to their electronic health records and laboratory test result data through patient portals, laboratory test results are often confusing and hard to understand. Many patients turn to web-based forums or question-and-answer (Q&A) sites to seek advice from their peers. The quality of answers from social Q&A sites on health-related questions varies significantly, and not all responses are accurate or reliable. Large language models (LLMs) such as ChatGPT have opened a promising avenue for patients to have their questions answered. OBJECTIVE: We aimed to assess the feasibility of using LLMs to generate relevant, accurate, helpful, and unharmful responses to laboratory test-related questions asked by patients and identify potential issues that can be mitigated using augmentation approaches. METHODS: We collected laboratory test result-related Q&A data from Yahoo! Answers and selected 53 Q&A pairs for this study. Using the LangChain framework and ChatGPT web portal, we generated responses to the 53 questions from 5 LLMs: GPT-4, GPT-3.5, LLaMA 2, MedAlpaca, and ORCA_mini. We assessed the similarity of their answers using standard Q&A similarity-based evaluation metrics, including Recall-Oriented Understudy for Gisting Evaluation, Bilingual Evaluation Understudy, Metric for Evaluation of Translation With Explicit Ordering, and Bidirectional Encoder Representations from Transformers Score. We used an LLM-based evaluator to judge whether a target model had higher quality in terms of relevance, correctness, helpfulness, and safety than the baseline model. We performed a manual evaluation with medical experts for all the responses to 7 selected questions on the same 4 aspects. RESULTS: Regarding the similarity of the responses from 4 LLMs; the GPT-4 output was used as the reference answer, the responses from GPT-3.5 were the most similar, followed by those from LLaMA 2, ORCA_mini, and MedAlpaca. Human answers from Yahoo data were scored the lowest and, thus, as the least similar to GPT-4-generated answers. The results of the win rate and medical expert evaluation both showed that GPT-4's responses achieved better scores than all the other LLM responses and human responses on all 4 aspects (relevance, correctness, helpfulness, and safety). LLM responses occasionally also suffered from lack of interpretation in one's medical context, incorrect statements, and lack of references. CONCLUSIONS: By evaluating LLMs in generating responses to patients' laboratory test result-related questions, we found that, compared to other 4 LLMs and human answers from a Q&A website, GPT-4's responses were more accurate, helpful, relevant, and safer. There were cases in which GPT-4 responses were inaccurate and not individualized. We identified a number of ways to improve the quality of LLM responses, including prompt engineering, prompt augmentation, retrieval-augmented generation, and response evaluation.

Development and preliminary validation of a MERS-CoV ELISA for serological testing of camels and alpacas.

This study describes the development and preliminary validation of a new serological assay using MERS-CoV S1 protein in an indirect enzyme-linked immunosorbent assay (ELISA) format. This assay has the advantage of being able to test MERS-CoV serum samples in a PC2 laboratory without the need for a high-level biocontainment laboratory (PC3 or PC4), which requires highly trained and skilled staff and a high level of resources and equipment. Furthermore, this MERS-CoV S1 ELISA enables a larger number of samples to be tested quickly, with results obtained in approximately five hours. The MERS-CoV S1 ELISA demonstrated high analytical specificity, with no cross-reactivity observed in serum of animals infected with other viruses, including different coronaviruses. We tested 166 positive and 40 negative camel serum samples and have estimated the diagnostic sensitivity (DSe) to be 99.4% (95% CI: 96.7 - 100.0%) and diagnostic specificity (DSp) to be 100% (95% CI: 97.2%-100.0%) relative to the assigned serology results (ppNT and VNT) using a S/P ratio cut-off value of >0.58. The findings of this study showed that our MERS-CoV S1 ELISA was more sensitive than the commercial EUROIMMUN ELISA (Se 99.4% vs 84.9%) and comparable to the ppNT assay, and therefore could be used as a diagnostic aid in countries in the Middle East where MERS-CoV is endemic in dromedary camels. The assay reagents and protocol were easily adapted and transferred from an Australian laboratory to a laboratory in the University of Hong Kong. Thus, the results described here show that the MERS-CoV S1 ELISA represents a cheap, rapid, robust, and reliable assay to support surveillance of MERS-CoV in camels in endemic regions.

The vicuna (Vicugna vicugna) as a natural host of Dictyocaulus filaria in Peru.

Lungworm infection, or verminous pneumonia, is a parasitic disease that causes serious problems in small and large ruminants. Despite the fact that nematodes of the genus Dictyocaulus in cattle and sheep are the main cause of this disease, there are few studies on the natural infections of South American camelids. For this reason, this study aims to report the natural infection by Dictyocaulus filaria in vicunas (Vicugna vicugna) for the first time. During a shearing season (chaku) in Cuzco, Peru, two accidentally killed adult vicunas were submitted to the IVITA-Marangani research center in Cuzco for their respective necropsies. The tracheas of both vicunas had numerous nematodes, as seen during the necropsy. The nematodes were collected in 70% ethanol and were morphologically identified as D. filaria. Likewise, the DNA of six nematodes was extracted, and the ITS2 region and the 28S rRNA gene were amplified and sequenced. The nucleotide sequences of both genetic markers were up to 100% identical with previously reported D. filaria DNA sequences found in the goat yearlings from Turkey, sheep from Iran, Turkey, and India, and the argali from Uzbekistan, which confirmed the morphological diagnosis. This finding represents the first molecular confirmation of a natural D. filaria infection in a South American camelid. It will be necessary to carry out future studies to know the current situation of verminous pneumonia in domestic and wild South American camelids and to know the negative effects of the disease on them.

Current safeguards, risk mitigation, and transparency measures of large language models against the generation of health disinformation: repeated cross sectional analysis.

OBJECTIVES: To evaluate the effectiveness of safeguards to prevent large language models (LLMs) from being misused to generate health disinformation, and to evaluate the transparency of artificial intelligence (AI) developers regarding their risk mitigation processes against observed vulnerabilities. DESIGN: Repeated cross sectional analysis. SETTING: Publicly accessible LLMs. METHODS: In a repeated cross sectional analysis, four LLMs (via chatbots/assistant interfaces) were evaluated: OpenAI's GPT-4 (via ChatGPT and Microsoft's Copilot), Google's PaLM 2 and newly released Gemini Pro (via Bard), Anthropic's Claude 2 (via Poe), and Meta's Llama 2 (via HuggingChat). In September 2023, these LLMs were prompted to generate health disinformation on two topics: sunscreen as a cause of skin cancer and the alkaline diet as a cancer cure. Jailbreaking techniques (ie, attempts to bypass safeguards) were evaluated if required. For LLMs with observed safeguarding vulnerabilities, the processes for reporting outputs of concern were audited. 12 weeks after initial investigations, the disinformation generation capabilities of the LLMs were re-evaluated to assess any subsequent improvements in safeguards. MAIN OUTCOME MEASURES: The main outcome measures were whether safeguards prevented the generation of health disinformation, and the transparency of risk mitigation processes against health disinformation. RESULTS: Claude 2 (via Poe) declined 130 prompts submitted across the two study timepoints requesting the generation of content claiming that sunscreen causes skin cancer or that the alkaline diet is a cure for cancer, even with jailbreaking attempts. GPT-4 (via Copilot) initially refused to generate health disinformation, even with jailbreaking attempts-although this was not the case at 12 weeks. In contrast, GPT-4 (via ChatGPT), PaLM 2/Gemini Pro (via Bard), and Llama 2 (via HuggingChat) consistently generated health disinformation blogs. In September 2023 evaluations, these LLMs facilitated the generation of 113 unique cancer disinformation blogs, totalling more than 40 000 words, without requiring jailbreaking attempts. The refusal rate across the evaluation timepoints for these LLMs was only 5% (7 of 150), and as prompted the LLM generated blogs incorporated attention grabbing titles, authentic looking (fake or fictional) references, fabricated testimonials from patients and clinicians, and they targeted diverse demographic groups. Although each LLM evaluated had mechanisms to report observed outputs of concern, the developers did not respond when observations of vulnerabilities were reported. CONCLUSIONS: This study found that although effective safeguards are feasible to prevent LLMs from being misused to generate health disinformation, they were inconsistently implemented. Furthermore, effective processes for reporting safeguard problems were lacking. Enhanced regulation, transparency, and routine auditing are required to help prevent LLMs from contributing to the mass generation of health disinformation.

Classifying alkaliphilic proteins using embeddings from protein language model.

Alkaliphilic proteins have great potential as biocatalysts in biotechnology, especially for enzyme engineering. Extensive research has focused on exploring the enzymatic potential of alkaliphiles and characterizing alkaliphilic proteins. However, the current method employed for identifying these proteins that requires web lab experiment is time-consuming, labor-intensive, and expensive. Therefore, the development of a computational method for alkaliphilic protein identification would be invaluable for protein engineering and design. In this study, we present a novel approach that uses embeddings from a protein language model called ESM-2(3B) in a deep learning framework to classify alkaliphilic and non-alkaliphilic proteins. To our knowledge, this is the first attempt to employ embeddings from a pre-trained protein language model to classify alkaliphilic protein. A reliable dataset comprising 1,002 alkaliphilic and 1,866 non-alkaliphilic proteins was constructed for training and testing the proposed model. The proposed model, dubbed ALPACA, achieves performance scores of 0.88, 0.84, and 0.75 for accuracy, f1-score, and Matthew correlation coefficient respectively on independent dataset. ALPACA is likely to serve as a valuable resource for exploring protein alkalinity and its role in protein design and engineering.

Advancing entity recognition in biomedicine via instruction tuning of large language models.

MOTIVATION: Large Language Models (LLMs) have the potential to revolutionize the field of Natural Language Processing, excelling not only in text generation and reasoning tasks but also in their ability for zero/few-shot learning, swiftly adapting to new tasks with minimal fine-tuning. LLMs have also demonstrated great promise in biomedical and healthcare applications. However, when it comes to Named Entity Recognition (NER), particularly within the biomedical domain, LLMs fall short of the effectiveness exhibited by fine-tuned domain-specific models. One key reason is that NER is typically conceptualized as a sequence labeling task, whereas LLMs are optimized for text generation and reasoning tasks. RESULTS: We developed an instruction-based learning paradigm that transforms biomedical NER from a sequence labeling task into a generation task. This paradigm is end-to-end and streamlines the training and evaluation process by automatically repurposing pre-existing biomedical NER datasets. We further developed BioNER-LLaMA using the proposed paradigm with LLaMA-7B as the foundational LLM. We conducted extensive testing on BioNER-LLaMA across three widely recognized biomedical NER datasets, consisting of entities related to diseases, chemicals, and genes. The results revealed that BioNER-LLaMA consistently achieved higher F1-scores ranging from 5% to 30% compared to the few-shot learning capabilities of GPT-4 on datasets with different biomedical entities. We show that a general-domain LLM can match the performance of rigorously fine-tuned PubMedBERT models and PMC-LLaMA, biomedical-specific language model. Our findings underscore the potential of our proposed paradigm in developing general-domain LLMs that can rival SOTA performances in multi-task, multi-domain scenarios in biomedical and health applications. AVAILABILITY AND IMPLEMENTATION: Datasets and other resources are available at https://github.com/BIDS-Xu-Lab/BioNER-LLaMA.

Rhododendron poisoning in alpacas (Vicugna pacos) in Northern Germany.

Poisoning is often suspected to be the origin of disease in South American camelids (SACs) by owners, but only in a few cases this assumption can be confirmed. In small ruminants, rhododendron poisoning is a common emergency for livestock veterinarians. However, this condition has rarely been reported in SACs so far. This paper provides information regarding clinical findings, hematology, clinical chemistry, and treatment of four alpacas after presumed intake of rhododendron leaves including pathological findings of one of the animals. Rhododendron leaves contain grayanatoxins that lead to hyperpolarization of excitable cells. Clinical signs that were observed in the presented alpacas comprised: salivation, dehydration, decreased motility of compartment 1, uncoordinated regurgitation, and cardiac arrhythmia. Clinical chemistry revealed that rhododendron poisoning was associated with metabolic acidosis and azotaemia, hyponatremia and hyperkalemia. Most striking macroscopic and histopathological findings included gastric ulceration, and renal infarcts along with inflammatory changes. Leaves of Rhododendron spp. were identified in the forestomach content of this animal. Affected animals were treated symptomatically as there is no specific antidote in rhododendron poisoning. This included parenteral rehydration, treatment of metabolic acidosis (infusion of sodium bicarbonate solution), and oral administration of activated charcoal to bind potential toxins. In addition, antibiotic treatment might be necessary to prevent aspiration pneumonia in case of uncoordinated regurgitation. Of the four animals, the worst affected alpaca was euthanized, one had minimal signs and two responded to supportive care and recovered. In conclusion, rhododendron poisoning might be fatal for alpacas in individual cases and therefore rhododendron bushes should not be placed in the habitat of SACs.

Histological evaluation of the alpaca (Vicugna pacos) vomeronasal organ.

Little is known about the neuronal structure of the vomeronasal organ (VNO), a receptor organ responsible for pheromone perception, in the alpaca (Vicugna pacos). This study was performed to determine the localization of neuronal elements, including protein gene product 9.5 (PGP 9.5), a pan-neuronal marker, olfactory marker protein (OMP), a marker of mature olfactory receptor cells, and phospholipase C beta 2 (PLC-ß2), a marker of solitary chemoreceptor cells (SCCs), in the VNO. OMP was identified in receptor cells of the vomeronasal sensory epithelium (VSE), while PGP 9.5 and PLC-ß2 were localized in both the VSE and vomeronasal non-sensory epithelium. Collectively, these results suggested that the alpaca VNO possesses SCCs and olfactory receptor cells, which recognize both harmful substances and pheromones.

Systematic analysis of ChatGPT, Google search and Llama 2 for clinical decision support tasks.

It is likely that individuals are turning to Large Language Models (LLMs) to seek health advice, much like searching for diagnoses on Google. We evaluate clinical accuracy of GPT-3·5 and GPT-4 for suggesting initial diagnosis, examination steps and treatment of 110 medical cases across diverse clinical disciplines. Moreover, two model configurations of the Llama 2 open source LLMs are assessed in a sub-study. For benchmarking the diagnostic task, we conduct a naïve Google search for comparison. Overall, GPT-4 performed best with superior performances over GPT-3·5 considering diagnosis and examination and superior performance over Google for diagnosis. Except for treatment, better performance on frequent vs rare diseases is evident for all three approaches. The sub-study indicates slightly lower performances for Llama models. In conclusion, the commercial LLMs show growing potential for medical question answering in two successive major releases. However, some weaknesses underscore the need for robust and regulated AI models in health care. Open source LLMs can be a viable option to address specific needs regarding data privacy and transparency of training.