6112 research outputs found
Sort by
XModNN: Explainable Modular Neural Network to Identify Clinical Parameters and Disease Biomarkers in Transcriptomic Datasets
Get PDFThe Explainable Modular Neural Network (XModNN) enables the identification of biomarkers, facilitating the classification of diseases and clinical parameters in transcriptomic datasets. The modules within XModNN represent specific pathways or genes of a functional hierarchy. The incorporation of biological insights into the architectural design reduced the number of parameters. This is further reinforced by the weighted multi-loss progressive training, which enables successful classification with a reduced number of replicates. The combination of this workflow with layer-wise relevance propagation ensures a robust post hoc explanation of the individual module contribution. Two use cases were employed to predict sex and neuroblastoma cell states, demonstrating that XModNN, in contrast to standard statistical approaches, results in a reduced number of candidate biomarkers. Moreover, the architecture enables the training on a limited number of examples, attaining the same performance and robustness as support vector machine and random forests. The integrated pathway relevance analysis improves a standard gene set overrepresentation analysis, which relies solely on gene assignment. Two crucial genes and three pathways were identified for sex classification, while 26 genes and six pathways are highly important to discriminate adrenergic–mesenchymal cell states in neuroblastoma cancer
Navigating the Spectrum: Assessing the Concordance of ML-Based AI Findings with Radiology in Chest X-Rays in Clinical Settings
Get PDFBackground
The integration of artificial intelligence (AI) into radiology aims to improve diagnostic accuracy and efficiency, particularly in settings with limited access to expert radiologists and in times of personnel shortage. However, challenges such as insufficient validation in actual real-world settings or automation bias should be addressed before implementing AI software in clinical routine.
Methods
This cross-sectional study in a maximum care hospital assesses the concordance between diagnoses made by a commercial AI-based software and conventional radiological methods augmented by AI for four major thoracic pathologies in chest X-ray: fracture, pleural effusion, pulmonary nodule and pneumonia. Chest radiographs of 1506 patients (median age 66 years, 56.5% men) consecutively obtained between January and August 2023 were re-evaluated by the AI software InferRead DR Chest ® .
Results
Overall, AI software detected thoracic pathologies more often than radiologists (18.5% vs. 11.1%). In detail, it detected fractures, pneumonia, and nodules more frequently than radiologists, while radiologists identified pleural effusions more often. Reliability was highest for pleural effusions (0.63, 95%-CI 0.58–0.69), indicating good agreement, and lowest for fractures (0.39, 95%-CI 0.32–0.45), indicating moderate agreement.
Conclusions
The tested software shows a high detection rate, particularly for fractures, pneumonia, and nodules, but hereby produces a nonnegligible number of false positives. Thus, AI-based software shows promise in enhancing diagnostic accuracy; however, cautious interpretation and human oversight remain crucial
Direct Effects of Poloxamer 188 on Mouse Brain Microvascular Endothelial Cells in an In Vitro Traumatic Brain Injury Model
Get PDFTraumatic brain injury (TBI) remains to be a major cause of disability and death, affecting millions of people each year worldwide. Threatening the survival of multiple cell types, including neurons and the vascular endothelial system, TBI compromises cellular membrane integrity, endangers blood-brain-barrier (BBB) function and cerebral blood flow resulting in ischemia and therefore amplifying cerebral injury. Not only the impact itself, injuring brain tissue instantly, but mostly secondary injury, originating from the impact and potentially lasting for years endanger brain tissue. Because the primary injury can only be prevented by precautions, most research aims to elucidate the complex mechanisms underlying secondary injury. Even though the comprehension of pathomechanisms underlying secondary injuries improves rapidly therapies specifically targeting cell death, oxidative stress, inflammation, BBB breakdown, excitotoxicity, and edema formation are not yet fully established. The vasculature is thought to be of major importance in brain injury, closely linked to disability. Saving the vascular endothelial system might improve clinical outcome after TBI.
Poloxamer 188 (P188) has been revealed to enhance neuronal function after ischemia/reperfusion (I/R) injury and after TBI in vitro. Simulations proposed that P188 might be able to protect the cellular membrane from damage by resealing membrane defects through insertion into pores. Additionally to protecting neuronal cell membranes, one mechanism aiding neuroprotection is thought to be shielding cells comprising the BBB. In this in vitro study, the direct effect of P188 treatment upon reoxygenation on brain endothelial cells was observed in simulated TBI, applying mild-to-moderate and severe damage. Mouse primary brain microvascular endothelial cells (MBEC) were exposed to normoxic or hypoxic conditions for either 5 or 15 h. 1 h compression was added, and P188 was administered during 2 h reoxygenation. A highly suitable in vitro compression-type TBI model, while simultaneously mimicking I/R injury and contusion, using brain endothelial cells was established by assessing cell number/viability, cytotoxicity/membrane damage, metabolic activity, apoptotic activity and total nitric oxide (NO) production.
While this study showed a direct protective effect of P188 in mild-to-moderate in vitro TBI, supporting the hypothesis, these effects were reduced after severe injury. P188, presented during reoxygenation, effected MBEC functioning directly. However, the effects differed between the assays performed, questioning the exact mechanism underlying endothelial protection. This study suggests that the protective properties of P188 on endothelial cells exceed the amphiphilic structure of P188, as some protective effect is shared with the osmotic control. In the future, the mechanism of action of P188, and especially the role of NO in neuroprotection, deserve to be further elucidated.Das Schädel-Hirn-Trauma (SHT) ist nach wie vor eine der Hauptursachen für Erwerbsunfähigkeit und Tod. Jährlich sind weltweit Millionen von Menschen betroffen; das Überleben unterschiedlicher Zelltypen, unter anderem von Neuronen und vaskulären Endothelzellen, ist bedroht. Das SHT beeinträchtigt die Integrität der Zellmembran und gefährdet die Funktion der Blut-Hirn-Schranke sowie des zerebralen Blutflusses. Als weitere Folge führt es zu Ischämie, wodurch die zerebrale Schädigung weiter verstärkt wird. Nicht nur das Trauma selbst, welches das zerebrale Gewebe augenblicklich schädigt, sondern vor allem sekundäre Schädigungen, die von dem Trauma ausgehen und möglicherweise jahrelang anhalten, schädigen das Hirngewebe. Da der Primärschaden nur durch Primärprävention verhindert werden kann, zielt ein Großteil der Forschung darauf ab, die komplexen Mechanismen der sekundären Schädigung aufzuklären. Obwohl sich das Verständnis der Pathomechanismen sekundärer Schädigungen wie Ischämie-Reperfusion rapide verbessert, fehlen spezifische Therapien, die Zelltod, oxidativen Stress, Inflammation, Blut-Hirn-Schranken Dysfunktion, Exzitotoxizität und Ödembildung verhindern. Es wird angenommen, dass das Gefäßsystem bei Hirnverletzungen von großer Bedeutung ist. Der Schutz des vaskulären Endothelsystems könnte das klinische Ergebnis nach einem SHT verbessern.
Studien haben gezeigt, dass Poloxamer 188 (P188) die neuronale Funktion nach Ischämie/Reperfusionsschaden und nach einem SHT in vitro verbessern kann. In Simulationen schützte P188 die Zellmembran vor Schädigung, indem es durch Membrandefekte entstandene Poren verschloss. Zusätzlich zum direkten Schutz von neuronalen Zellmembranen könnte ein Mechanismus der Neuroprotektion die Zellen der Blut-Hirn-Schranke erfassen. In dieser Arbeit wurde untersucht, ob die Behandlung mit P188 nach einem in vitro simulierten SHT mit einerseits milder bis moderater und andererseits schwerer Intensität während der Reoxygenierungsphase einen direkten Effekt auf Endothelzellen des Gehirns hat.
Hierzu wurde ein geeignetes in vitro SHT-Modell mit primären murinen mikrovaskulären Endothelzellen des Gehirns entwickelt. Die Zellen wurden normoxischen oder hypoxischen Bedingungen für 5 und für 15 Stunden ausgesetzt. Zudem wurden die Zellen für eine Stunde simultan komprimiert. Anschließend wurden sie während der Reoxygenierungsphase für 2 Stunden mit P188 behandelt. Ein direkter Effekt von P188 auf zerebrale Endothelzellen wurde über Zellzahl/Viabilität, Zytotoxizität/Membranschädigung, metabolische Aktivität, Produktion von Stickstoffmonoxid (NO) sowie apoptotische Aktivität untersucht. P188 verbesserte einerseits die Zellzahl/Viabilität und die metabolische Aktivität und andererseits wurde die Zytotoxizität/der Membranschaden nach mildem-moderatem SHT reduziert. Nach stark ausgeprägter Schädigung war dieser Effekt von P188 deutlich reduziert.
Zusammenfassend konnte die Hypothese eines direkten Effekts von P188 auf primäre murine mikrovaskuläre Endothelzellen des Gehirns gestützt werden. Weitere Studien, in denen der genaue Wirkmechanismus von P188, insbesondere im Blick auf die NO-Produktion, weiterführend erforscht wird, sind erforderlich
Effekte okklusaler Belastung auf die Progression über 16 Jahre von Nichtkariösen Zahnhalsdefekten – Ergebnisse der „Study of Health in Pomerania“ (SHIP)
Get PDFAktuelle wissenschaftliche Forschungen diskutieren weiterhin einen möglichen Zusammenhang zwischen okklusaler Belastung (OW) und der Progression von Nichtkariösen Zahnhalsdefekten (NCCLs). Der bisherige Stand der Untersuchungen beschreibt unterschiedliche Möglichkeiten von Entstehungstheorien von NCCLs. Unter anderem, dass durch OW entstehende Spannungen in der zervikalen Schmelzmatrix Absprengungen in dieser Region auftreten. Viele Untersuchungen bestärken eine multifaktorielle Entstehungstheorie für NCCLs. Hier spielen mechanische, wie zum Beispiel Bruxismus als Parafunktion oder Abrieb durch Zähneputzen, aber auch chemische und genetische Aspekte eine Rolle. Chemisch wird die Zahnhartsubstanz vor allem durch Säureeinfluss verändert, was unter anderem durch häufige Zufuhr von Frucht- und Gemüsesäften oder Softdrinks, aber auch anderen Ernährungsgewohnheiten auftritt. Die nicht eindeutig geklärte Lage zur Entstehung von NCCLs und deren Zusammenhang erfordert weitere Untersuchungen zu diesem Thema.
Aus diesem Grund hat sich die vorliegende Arbeit die Frage zu Grunde gelegt, ob OW einen Einfluss auf die Entstehung von NCCLs hat. Daten aus der populationsbasierten Kohortenstudie Study of Health in Pomerania (SHIP) ermöglichten einen longitudinalen Studienaufbau über 16 Jahre. Im Rahmen dieser Studie wurde mit den Probanden eine Anamnese durchgeführt. Genaue Befragung bezüglich des Lebens- und Ernährungsstils wurden ergänzt. Darüber hinaus wurden Abformungen und Gipsmodelle angefertigt.
In unserer Untersuchung wurden die Größe der NCCL- sowie der OW-Flächen an digitalisierten Modellen des Ober- und Unterkiefers von 364 Studienteilnehmern vermessen.
Es konnten im Quer- und Längsschnitt statistische Modellkombinationen hinsichtlich der Angewohnheiten der Probanden erstellt werden: unter anderem wurde an Alter, Geschlecht, Zahnputzangewohnheiten (zweimal oder häufiger pro Tag), Frucht- und Gemüsesaftkonsum und Sodbrennen angepasst.
So konnte in der Querschnittanalyse ein signifikanter Zusammenhang zwischen dem Auftreten von OW und NCCLs festgestellt werden. Die Ergebnisse des longitudinalen Vergleichs zeigten eine ähnliche Tendenz, wenn auch nicht signifikant. Trotzdem sind weiterhin Studien zu dieser Thematik erforderlich, welche eventuell einen größeren Umfang haben und weitere eventuell relevante Faktoren einbeziehen.Background: It is still discussed whether occlusal wear (OW) affects the formation of non-carious cervical lesions (NCCLs).
Objective: To estimate effects of OW on the presence and development of NCCLs, using 16-year follow-up data from a cohort study.
Methods: Occlusal and cervical defects were measured in 728 cast models (one from the upper jaw and one from the lower jaw) of 364 participants. Adjusted mixed-effects ordinal logistic models analysing estimated cross-sectional (N = 1308 teeth/291 subjects) and longitudinal (N = 718 teeth/226 subjects) associations of OW with NCCLs using tooth level data.
Results: OW size was cross-sectionally (OR = 1.74; 95% CI: 1.27–2.38 for OW size; OR = 0.97; 95% CI: 0.94–0.99 for squared OW size), but not longitudinally (OR = 1.14; 95% CI: 0.99–1.30) associated with odds of higher NCCL sizes. For cross-sectional analyses, predicted probabilities of an NCCL size of 0 decreased from about 0.996 to 0.010 for OW sizes of 0 to 25.
Conclusion: Results suggest an association between OW and NCCL size. However, as longitudinal results were non-significant, while consistent in direction, large-scaled cohort studies are demanded to more precisely estimate effect strength.
K E Y WO R D S
cohort study, cross-sectional study, epidemiology, longitudinal study, non-carious cervical lesions, occlusal wea
Bat white-nose disease fungus diversity in time and space
Get PDFWhite-nose disease (WND), caused by the psychrophilic fungus Pseudogymnoascus destructans, represents one of the greatest threats for North American hibernating bats. Research on molecular data has significantly advanced our knowledge of various aspects of the disease, yet more studies are needed regarding patterns of P. destructans genetic diversity distribution. In the present study, we investigate three sites within the native range of the fungus in detail: two natural hibernacula (karst caves) in Bulgaria, south-eastern Europe and one artificial hibernaculum (disused cellar) in Germany, northern Europe, where we conducted intensive surveys between 2014 and 2019. Using 18 microsatellite and two mating type markers, we describe how P. destructans genetic diversity is distributed between and within sites, the latter including differentiation across years and seasons of sampling; across sampling locations within the site; and between bats and hibernaculum walls. We found significant genetic differentiation between hibernacula, but we could not detect any significant differentiation within hibernacula, based on the variables examined. This indicates that most of the pathogen’s movement occurs within sites. Genotypic richness of P. destructans varied between sites within the same order of magnitude, being approximately two times higher in the natural caves (Bulgaria) compared to the disused cellar (Germany). Within all sites, the pathogen’s genotypic richness was higher in samples collected from hibernaculum walls than in samples collected from bats, which corresponds with the hypothesis that hibernacula walls represent the environmental reservoir of the fungus. Multiple pathogen genotypes were commonly isolated from a single bat (i.e. from the same swab sample) in all study sites, which might be important to consider when studying disease progression
Prediction of subcutaneous drug absorption − Development of novel simulated interstitial fluid media for predictive subcutaneous in vitro assays
Get PDFMedia that mimic physiological fluids at the site of administration have proven to be valuable in vitro tools for predicting in vivo drug release, particularly for routes of administration where animal studies cannot accurately predict human performance. The objective of the present study was to develop simulated interstitial fluids (SISFs) that mimic the major components and physicochemical properties of subcutaneous interstitial fluids (ISFs) from preclinical species and humans, but that can be easily prepared in the laboratory and used in in vitro experiments to estimate in vivo drug release and absorption of subcutaneously administered formulations. Based on data from a previous characterization study of ISFs from different species, two media were developed: a simulated mouse-rat ISF and a simulated human-monkey ISF. The novel SISFs were used in initial in vitro diffusion studies with a commercial injectable preparation of liraglutide. Although the in vitro model used for this purpose still requires significant refinement, these two new media will undoubtedly contribute to a better understanding of the in vivo performance of subcutaneous injectables in different species and will help to reduce the number of unnecessary in vivo experiments in preclinical species by implementation in predictive in vitro models
The Hidden Burden of Healthcare Financing and its Inequality in Cambodia: Revealing the reality beyond the Sustainable Development Goals’ indicators.
Get PDFBackground: Universal Health Coverage (UHC) aims for equitable access to healthcare, minimizing the financial burden (FB) on households. This dissertation examines the effectiveness of standard indicators like impoverishment and catastrophic health expenditure (CHE) in reflecting FB and its inequality in Cambodia. Recent publications have highlighted Cambodia’s progress in reducing CHE, impoverishment, and inequality. However, this dissertation's core contention is that by using the standard metrics of FB, one masks its impact among the lowest-income groups and inequality, misrepresenting trends and policy achievements.
Objectives: This dissertation evaluates FB using standard and alternative measures to understand FB trends and inequalities across Cambodia's population, and appropriately reflect the burden on lowest-income households, thus better informing policy responses for achieving UHC.
Methods: Employing data from the Cambodian Socio-Economic Surveys (CSES) from 2004 to 2019/20, this dissertation comprises three studies. The first assesses standard FB indicators and resulting inequality across socioeconomic and regional groups. The second examines impoverishment sensitivity when building various consumption aggregates and setting poverty lines. The third introduces an alternative FB measure, Excessive Financial Burden (EFB), comparing it against standard CHE metrics to evaluate financial hardship and associated inequality.
Results: Initial findings showed a 27% decrease in standard CHE incidence from 2004 to 2009, with significant reductions in FB for households seeking medical care. However, from 2009 onwards, CHE reductions stagnated. The sensitivity analysis revealed that methodological choices highly influence impoverishment estimates, making it a poor choice to assess FB trends. The introduction of EFB highlighted worsening FB from 2009 to 2019, particularly among the lowest-income households, with EFB inequality deepening significantly. EFB with wealth index rankings appears more effective at capturing economic shocks due to healthcare spending, and its decomposition provided valuable insights to mitigate inequality in the medium term.
Conclusions: The study underscores the limitations of standard CHE and impoverishment measures in capturing the full extent of FB and suggests that alternative measures like EFB, combined with wealth indices, provide a more comprehensive view of economic shocks due to healthcare spending. The findings advocate for policy measures to extend social health protection and improve targeting to address FB and inequality effectively. The methodology and insights from this study offer a framework for other low and middle-income countries striving to achieve UHC.Hintergrund: Universelle Gesundheitsversorgung (UHC) zielt auf einen gleichberechtigten Zugang zu
Gesundheitsdienstleistungen ab und minimiert die finanzielle Belastung der Haushalte (FB). Diese
Dissertation untersucht die Wirksamkeit von Standardindikatoren wie Verarmung und katastrophale
Gesundheitsausgaben (CHE) bei der Betrachtung von FB und ihrer Ungleichheit in Kambodscha.
Jüngste Veröffentlichungen haben Kambodschas Fortschritte bei der Reduzierung von CHE,
Verarmung und Ungleichheit bekräftigt. Die Kernaussage dieser Dissertation ist jedoch, dass man
durch die Verwendung der Standardmetriken von FB die Auswirkungen auf die
einkommensschwächsten Gruppen verschleiert und politische Errungenschaften falsch darstellt.
Ziele: Diese Dissertation evaluiert FB unter Verwendung von Standard- und alternativen Messungen,
um FB-Trends und Ungleichheiten in der kambodschanischen Bevölkerung zu verstehen und die
Belastung der einkommensschwächsten Haushalte angemessen widerzuspiegeln, um besser
geeignete politische Maßnahmen formulieren zu können.
Methodik: Unter Verwendung von Daten aus den kambodschanischen sozioökonomischen
Erhebungen (CSES) von 2004 bis 2019/20 umfasst diese Dissertation drei Studien. Die erste bewertet
Standard-FB-Indikatoren und die daraus resultierende Ungleichheit zwischen sozioökonomischen
und regionalen Gruppen. Die zweite Studie untersucht die Sensitivität des Verarmungindikators bei
der Bildung verschiedener Konsumaggregate und der Festlegung von Armutsgrenzen. Die dritte stellt
eine alternative FB-Messgröße, die übermäßige finanzielle Belastung (EFB), vor und vergleicht sie mit
Standard-CHE-Metriken zur Bewertung von FB und damit verbundener Ungleichheit.
Ergebnisse: Erste Befunde zeigten einen Rückgang der Standard-CHE-Inzidenz um 27% von 2004 bis
2009, mit einer signifikanten Reduktion der FB in Haushalten, die medizinische Versorgung in
Anspruch nahmen. Die Sensitivitätsanalyse ergab, dass kömmlischen methodische Entscheidungen
einen großen Einfluss auf die Schätzungen der Verarmung haben. Die Einführung des EFB hat gezeigt,
dass sich die FB und seine Ungleicheit von 2009 bis 2019 verschlechtert hat, insbesondere bei den
Haushalten mit dem niedrigsten Einkommen.
Schlussfolgerungen: Diese Studie unterstreichen die Grenzen von standard Metriken bei der
Erfassung des vollen Ausmaßes von FB und legt nahe, dass alternative Messgrößen wie EFB in
Kombination mit Vermögensindizes einen umfassenderen Überblick über wirtschaftliche Schocks
aufgrund von Gesundheitsausgaben bieten. Die Methodik und die Erkenntnisse aus dieser Studie
bieten einen Rahmen für andere Länder mit niedrigem und mittlerem Einkommen, die eine
universelle Gesundheitsversorgung anstreben. Die Ergebnisse plädieren für dringende politische
Maßnahmen zur Ausweitung und Überarbeitung des sozialen Gesundheitsschutzes in Kambodscha
The role of frontal cortex in novel-word learning and consolidation: Evidence from focal transcranial direct current stimulation
Get PDFPrevious studies have demonstrated that conventional transcranial direct current stimulation (tDCS) can enhance novel-word learning. However, because of the widespread current that is induced by these setups and lack of appropriate control conditions, little is known about the underlying neural mechanisms. In the present double-blinded and sham-tDCS controlled study, we investigated for the first time if regionally precise focal tDCS targeting two key nodes of the novel-word learning network at different time points would result in regionally and temporally distinct effects. 156 participants completed a contextual novel-word-learning paradigm and learning success was probed immediately after the acquisition period and 30-min later. Participants were randomly assigned to six stimulation conditions: Active tDCS (1.5 mA) was administered to left inferior frontal (IFG) or middle temporal gyrus (MTG), either during acquisition or delayed recall. Control groups received sham-tDCS either during acquisition or delayed recall (50% IFG/MTG). Data were analyzed with a generalized linear mixed model with a binomial link function in a Bayesian framework. Our results showed that frontal tDCS selectively increased accuracy gains from immediate to delayed recall, irrespective of timing of the stimulation. There was no evidence for beneficial effects of middle temporal gyrus tDCS. Our findings confirm that IFG tDCS can enhance novel-word learning in a regionally, but not timing specific way. Tentatively, this may be explained by enhancement of semantic selection processes resulting in more effective consolidation and/or retrieval. Future studies using longer time intervals between assessments are required to clarify the potential contribution of neurophysiological after-effects of IFG tDCS administered during acquisition to enhanced consolidation
Solving algebraic equations by using the bar model: Theoretical and empirical considerations
Get PDFSolving equations is known to bear several challenges for learners. We discuss an approach based on conceptual understanding regarding the transformation of equations with the help of the so-called bar model in combination with the transposing strategy. First, we sketch shortly the main ideas that guided the development of the learning environment. Second, we discuss insights from the first design experiments with six students working with equation transformation in their regular school curriculum. These design experiments are embedded in a design research approach. In particular, we zoom into the semiotic processes of how learners connect several representations and emphasize a varying difficulty regarding single concept elements necessary to understand the concept of equivalent equations as a whole. Based on that, obstacles that come along with using the bar model are highlighted. Finally, we point to theoretical insights and implications for enhancing our learning environment
Development of Simulated Subcutaneous Interstitial Fluids for Predictive In vitro Assays Based on Cross-Species Data
Get PDFSC administration is used for delivering drugs which cannot be administered via the oral route. Yet, the absorption profile after SC administration is difficult to predict from preclinical studies. One of the main reasons for this poor prediction may be the differences in the physiology that preclinical species and humans have, but there is a lack of information about the SC tissue composition and, particularly, about the fluid that perfuses this area, the ISF. Better knowledge on the composition and physicochemical properties of the ISF may provide better insights into in vivo drug performance and even allow for developing simulated ISF to be used in in vitro models, to further investigate in vivo drug performance. To date, a comparison of ISF from different preclinical species and humans has not been published. Furthermore, few simulated ISF media have been proposed. Yet, those are literature-based and most of them cannot be regarded as biorelevant since non-physiological components are part of the composition. In this context, this study aimed to investigate the composition and physicochemical properties of the ISF from preclinical species and humans, to discern interspecies differences and, ultimately, to develop SISFs for in vitro predictive assays.
As a first step, ISF was isolated from the SC tissue of preclinical species and humans using the centrifugation method. This technique results in undisturbed ISF, a critical factor to consider given the invasiveness of the isolation process. Secondly, the isolated ISF samples were characterized using the same techniques across all samples, to enable data comparison. The characterization studies constituted a considerable contribution towards the understanding of the ISF composition across species. Similar values for pH, electrolyte content and osmolality were found for the ISF samples from the different species, while significant differences were observed for some parameters, particularly for albumin content. For instance, non-human primates ISF samples showed significantly higher albumin content than rodents ISF samples, while being in the same range than human ISF samples.
After the isolation and characterization of ISF samples, the resulting data were utilized to develop simulated ISF media. These SISFs aimed to replicate as closely as possible, the experimental characterization data, while keeping overall composition and preparation of the media simple to ensure easy implementation in the laboratory. pH, osmolality, and electrolyte and albumin content were set as target parameters. Meanwhile, buffer capacity, colloid osmotic pressure and surface tension were defined as reference parameters, i.e., were only assessed after preparation. Initially, a buffer was prepared, considering the target pH, electrolyte content, and osmolality, as these parameters were consistent across all species. The physicochemical properties of this buffer were measured, and the buffer was designated as Blank SISF. Different albumin concentrations were incorporated into the Blank SISF for representing ISF of mice and rats (17.5 g/L) and non-human primates and humans (30.0 g/L), developing in this way two media: the SMR-ISF and the SHM-ISF. After preparation, the SISFs were screened in the same manner as the ISF samples, and the resulting data were compared. At this step, slight adjustments were made in the electrolyte composition to better approximate the desired physicochemical properties.
Finally, in vitro experiments with the SISFs were conducted to analyze the influence of their composition, particularly their albumin content, on drug diffusion from the interstitial to the intravascular space. Liraglutide was selected as the model drug due to its high affinity for albumin. For these experiments, the Artificial Subcutaneous Tissue assay was utilized. The latter consists of a transwell plate with a gel-like artificial matrix mimicking key components of the extracellular matrix, such as collagen and hyaluronic acid. For these studies, SISF was added on the matrix prior to the incorporation of the formulation to be tested and PBS was used as receiver medium. Since PBS does not contain albumin, it was defined as Blank SPlasma. In a second step, Blank SPlasma was spiked with albumin to mimic the composition of the plasma samples determined in the characterization studies (25 g/L for the SMR-Plasma and 60 g/L for the SHM-Plasma). In this way, the SMR-ISF was used in conjunction with the SMR-Plasma in one setup, and the SHM-ISF was used together with the SHM-Plasma in a second setup. Higher diffusion rates were observed when the respective SISF at the apical side was combined with Blank SPlasma at the basolateral side, i.e., in absence of albumin in the receiver media. The samples analysis revealed the diffusion of both liraglutide and albumin through the membrane until equilibrium was achieved. Therefore, the results obtained with Blank SPlasma may be more representative of the physiological diffusion of liraglutide than those obtained with SPlasma containing albumin. Overall, these new media are promising tools in the context of developing new biorelevant models aiming to explain the differences in the in vivo performance of SC injectables in preclinical species and humans. This is because biorelevant components such as albumin significantly impact the physicochemical properties of the media. However, other aspects, such as the fat content difference in the SC tissue or even the exact injection point, may also influence in vivo performance and could be considered for future investigations.
In conclusion, this study has provided essential insights into the composition and properties of the SC ISF, clarifying previously limited data. The findings suggest that, among the studied preclinical species, the ISF of non-human primates is the most similar to human ISF. Yet, further research is needed to understand which differences between non-human primates and humans influence the bioavailability profile disparities. The ISF characterization data were utilized to develop SISFs. The addition of albumin at targeted concentrations resulted in physicochemical properties closer to those of the reference ISF data. The results presented here indicate that these media offer several advantages over those previously described in literature. Principally, the SISFs provide a better representation of the ISF composition and physicochemical properties, while also accounting for the differences among species. Although the in vitro model used for the first in vitro experiments still requires refinement, the SISFs will undoubtedly help to better understand the in vivo performance of subcutaneously injected formulations. Furthermore, the SISFs can provide insights into how results obtained in preclinical species can be extrapolated to humans and, more importantly, avoid unnecessary animal testing. Overall, the work presented in this research therefore marks a significant step forward in developing in vitro tools to investigate the behavior of SC injectables and should inspire further investigation and refinement for future applications