Exploring multiple monitoring modalities for large-scale 3D tissue cultivation in bioreactor

Exploring multiple monitoring modalities for large-scale 3D tissue cultivation in bioreactor | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Exploring multiple monitoring modalities for large-scale 3D tissue cultivation in bioreactor Laura Chastagnier, Sarah Pragnere, Yilbert Gimenez, Celine Loubière, and 7 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7216794/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 07 Jan, 2026 Read the published version in Scientific Reports → Version 1 posted 11 You are reading this latest preprint version Abstract As the field of tissue engineering advances toward clinically viable, large-scale biofabrication, there is an urgent need for non-invasive, real-time monitoring tools capable of assessing the dynamic maturation of 3D bioprinted tissues. This study presents a modular analytical framework combining physicochemical, metabolic, morphological, and perfusion monitoring strategies tailored to volumetric engineered tissues. A perfused cultivation platform was developed for 10.8 cm³ bioprinted fibroblast tissues, enabling fine regulation of pH, temperature, and oxygen. Enhanced oxygen control was achieved through dual-gas PID regulation, reducing deviation from 128–22%. Metabolic activity was monitored via online Raman spectroscopy, allowing real-time lactic acid quantification with a prediction precision error of 0.103 g.L⁻¹, despite low secretion levels typical of adherent cells. Morphological evolution was tracked using 7 Tesla MRI, revealing high fidelity to initial designs (87.6% within 1 mm deviation) and providing longitudinal insights into tissue remodeling without labeling or sectioning. Perfusion was evaluated through computational fluid dynamics (CFD) simulations and MRI velocimetry, confirming flow heterogeneity and validating internal fluid distribution. These combined approaches demonstrate the feasibility of a closed-loop, feedback-driven biomanufacturing process that aligns with quality-by-design principles and regulatory expectations for advanced therapy medicinal products (ATMPs). The integration of established tools from pharmaceutical and clinical fields into tissue engineering workflows marks a critical step toward scalable, standardized, and adaptive biofabrication processes capable of supporting the next generation of functional tissue substitutes. Biological sciences/Biological techniques Biological sciences/Biotechnology Physical sciences/Engineering Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction The field of engineered tissue biofabrication is undergoing a transformative evolution, rapidly progressing toward the fabrication of fully functional, implantable human organs. This progress is driven by the convergence of advances in 3D bioprinting, stem cell engineering, biomaterials science, and bioinformatics. Today, several complementary techniques are being developed to reconstruct tissue architecture and function, including extrusion-based bioprinting 1 , 2 , digital light processing (DLP) 3 , stereolithography 4 , freeform reversible embedding of suspended hydrogels (FRESH) 5 , and volumetric bioprinting 6 . Extrusion-based bioprinting remains the most widely used due to its versatility and compatibility with a broad range of bioinks 7 . It has been applied to fabricate cartilage 8 , liver-like tissues 9 , and vascular networks with high precision 10 . More recently, volumetric bioprinting has emerged as a breakthrough technology, enabling the rapid fabrication of centimeter-scale soft tissues within seconds through tomographic light projections, achieving high spatial fidelity and cellular viability 11 The FRESH technique—where bioinks are printed into a gelatin microparticle support bath—has enabled the successful printing of complex, soft, and delicate structures such as collagen-based heart valves and mini hearts, with preserved geometry and mechanical properties 12 . Additionally, researchers have used DLP-based bioprinters to construct multiscale vasculature capable of supporting perfusable organoids, a significant step toward the vascularization of thick tissues 13 . These advancements are further supported by the integration of patient-specific imaging data and machine learning algorithms that allow personalized of biofabricated engineered tissues 14 . Coupled with the increasing maturity of stem cell-derived organoids and tissue-specific progenitor cells 15 , 16 , these technologies collectively bring the field closer to the production of clinically relevant, transplantable organs. However, as the complexity and dynamism of engineered tissues grow, traditional analytical methods are reaching their limitations. Most current approaches rely on destructive assays or endpoint analyses, which are incompatible with dynamic, iterative, and long-term biomanufacturing processes 17 . To ensure quality and functionality throughout tissue maturation, the field requires an integrated, non-invasive monitoring environment capable of providing real-time feedback during and after fabrication 18 . To facilitate this paradigm shift, a tailored analytical framework must be developed to support next-generation biofabrication workflows. This framework should be capable of capturing and interpreting the most critical parameters that define the maturation and performance of engineered tissues. Among these, four classes of parameters are of particular importance: physicochemical environment, metabolic activity, morphology and perfusion. 1. Physicochemical Environment. The local tissue environment—including pH, oxygen tension, osmolarity, and temperature—is a key regulator of cell behavior, influencing proliferation, differentiation, and apoptosis. Deviation from optimal conditions can impair tissue function and lead to cellular stress. Embeddable sensors capable of real-time, non-invasive tracking of these physicochemical parameters, including optical and electrochemical sensors for pH and oxygen, have been successfully integrated into bioreactor 19 , 20 . These tools allow the reconstruction of physiological gradients essential for mimicking in vivo -like tissue environments 21 , such as metabolic zonation in liver engineered tissues. 2. Metabolic Activity. Metabolic readouts provide a functional assessment of cellular health and activity 22 . Parameters such as glucose uptake, lactate production, or mitochondrial respiration reflect the bioenergetic state of cells within the engineered tissue. These metrics are particularly useful for evaluating differentiation status or the response to environmental changes. For instance, Raman spectroscopy has been used to monitor metabolic shifts 23 associated with stem cell differentiation without the need for destructive sampling 24 . Real-time metabolic monitoring can guide media exchange, differentiation protocols, and overall culture optimization 25 . 3. Morphology. Tissue morphology, encompassing both micro- and macro-architectural organization, is foundational to the structural integrity and functional identity of engineered tissues 26 . Proper spatial arrangement of cells and extracellular matrix is vital for enabling tissue-specific functions such as nutrient diffusion, signal propagation, and mechanical resistance. Non-invasive imaging techniques like magnetic resonance imaging (MRI) 27 and optical coherence tomography (OCT) 28 have been successfully applied to monitor morphology over time in 3D scaffolds and hydrogels 29 , 30 . 4. Perfusion. Adequate perfusion is essential to supply cells with oxygen and nutrients while removing metabolic waste products. In thick or vascularized engineered tissues, insufficient perfusion can lead to hypoxic zones and cell death, compromising the viability and function of the entire tissue 31 . Monitoring perfusion helps assess the effectiveness of biofabricated vasculature or perfusion bioreactors and is critical in scaling engineered tissues toward clinical dimensions 32 . To address the pressing need for real-time, non-invasive monitoring of biofabricated engineered tissues, the present study investigates a suite of integrated, non-destructive tools designed to assess critical parameters during tissue development ( Fig. 1 ). Specifically, we explore the monitoring and qualification of engineered tissues with a focus on the control of physicochemical conditions, compositional dynamics, and structural evolution over time. Our approach leverages a combination of advanced sensing and imaging technologies—including embedded temperature, pH and oxygen sensors, magnetic resonance imaging (MRI), and vibrational spectroscopy (e.g., Raman analysis)—to assess their respective and complementary contributions to a robust analytical framework. These tools are evaluated not only for their technical feasibility within a biofabrication context but also for their ability to support dynamic, adaptive tissue engineering workflows. By integrating these modalities into the biofabrication pipeline, we aim to demonstrate the feasibility of establishing a real-time, feedback-driven system for tissue quality assessment—an essential component for future scalable and clinically relevant tissue regeneration strategies. RESULTS Physicochemical Environment Monitoring in Engineered Tissues Tissue maturation protocols are typically developed on small-scale engineered tissues, often limited to the millimetre-cube (mm³) range, within partially controlled physicochemical environments. These studies are generally conducted in incubators with regulated temperature and buffered pH, yet oxygen delivery remains unregulated—despite its pivotal role in cell physiology. Such limited control becomes a significant bottleneck when attempting to scale protocols to larger engineered tissues in the centimetre-cube (cm³) range. Oxygen, in particular, plays a critical role in tissue development and regeneration, as reviewed by Zoneff et al. 21 . During the proliferative phase, oxygen is required for glycoprotein biosynthesis, while in the maturation and remodelling phases, it acts as a cofactor in hydroxyproline synthesis, essential for collagen fibril formation 33 . Despite this, reproducing native oxygen gradients in engineered tissues remains a major challenge. Furthermore, cellular oxygen demand varies significantly across cell types 34 . For instance, hepatocytes consume oxygen at rates of approximately 0.3 nmol/s/10⁶ cells, nearly five times higher than fibroblasts (0.05 nmol/s/10⁶ cells), due to their elevated metabolic activity 35 . This demand also fluctuates depending on cell cycle stage (e.g., stromal vs. terminally differentiated) and cell density within the scaffold. These complex dynamics, combined with the effects of both hypoxia and hyperoxia, underline the necessity for precise oxygen regulation. Notably, short-term hyperoxic pulses can induce protective effects—a phenomenon known as ischemic preconditioning—although prolonged exposure can be cytotoxic, underscoring the need for a narrow regulatory window. Our Approach: Scaling Physicochemical Control to 3D Bioprinted Engineered tissues To address these challenges, we implemented controlled cultivation protocols for 3D bioprinted engineered tissues with volumes up to 10.8 cm³ (2×2×2.7 cm; 12x12 pores). We applied our strategy to a conjunctive tissue composed of human fibroblast 36 and developed generic, robust regulation protocols. A special set-up was here necessary to enable the constant perfusion of the bioprinted engineered tissue with a regulated and monitored cultured media. The experimental system, depicted in Figure 2-A , feature a commercial regulation vessel equipped with temperature, pH, O 2 and Raman probes, tow gas inlets and a perfusion loop. The 10.8 cm³ tissue was placed in a specially design 3D printed perfusion vessel. Cultures were maintained for 6 days. Initial environmental control strategies were adapted from pharmaceutical bioprocessing protocols useful for cell suspension or microcarrier-based cultures ( Fig. 2-B ) 37 . Temperature and pH were effectively regulated using this initial protocol, with deviations from the set point of only 0.94% and 0.16%, respectively. In contrast, oxygen regulation proved more challenging (128% deviation from the set point) and required substantial optimisation to produce stable dissolved oxygen level in the perfused culture media. This steady and controlled level is mandatory to overcome the reduced growth rates of fibroblasts (doubling time of 3.5 days in 3D culture, compared to 1.7 days in 2D monolayers) we already demonstrate in bioprinted engineered tissues 36 . Initial regulation relied solely on positive oxygen control (air/O₂ injection in the gas inlet sparger), which proved insufficient since oxygen levels never reached the expected value (set point 35%). To improve regulation, a negative feedback loop using nitrogen (N₂) injection was introduced in the gas inlet sparger, enabling effective oxygen depletion. We further refined control using cascaded PID loop with optimised parameters. This involved a doubling of the proportional term (kP) and a 5-fold decrease of the Integral term (kD), reducing the dissolved oxygen level deviation from 128% to 22% ( Fig.2-C and Supplementary Table 1 ). These changes in the regulation cascade and PID values had an undesirable effect on the pH regulation. This effect was expected since the negative control of the pH is bound to the dissolved CO 2 concentration, which is injected, together with Air and N 2 in the gas mix. The negative oxygen control action was then accompanied by a concomitant negative control action on the pH. Finally, when the set points for pH and dissolved oxygen were set to 7.4 and 40% respectively, both controls were fully successful, with a dissolved oxygen level never lower than the 40% set point and a stable pH of 7.4. Insights and Implications for Bioprocess Control Our results also revealed clear monitoring of oxygen consumption. Indeed, bioprinted conjunctive tissue populated with human fibroblasts showed pronounced oscillations in dissolved oxygen (dO₂) ( Fig. 2-D ). This aligns with industrial cell culture processes, where dO₂ oscillations are used to estimate oxygen uptake rates (OUR) and reflect both cell number and oxygen consumption rates 38 . This finding suggests that dO₂ fluctuation patterns could serve as a real-time, non-invasive proxy for estimating cell number and metabolic activity within 3D engineered tissues—offering, to our knowledge, the first online qualitative description of the respiratory behaviour of tissue-like structures. Moreover, our monitoring system proved highly sensitive to external perturbations. As illustrated in Figure 2 (arrows), media additions and sample withdrawals generated identifiable deviations in the oxygen and pH regulation profiles. Such traceable events provide valuable information on the impact of interventions on the tissue’s microenvironment and could inform adaptive feedback in automated culture protocols—mirroring practices in pharmaceutical manufacturing. Finally, this ability to continuously log and interpret real-time data from the tissue environment is a key step toward creating regulatory-compliant production systems for advanced therapy medicinal products (ATMPs). It establishes a foundational layer for quality-by-design (QbD) approaches in tissue engineering 39-41 . Metabolic Activity Monitoring of Engineered Tissues Metabolic activity is a core indicator of tissue viability and functionality. Among the many metabolic markers available, lactic acid stands out for its dual relevance: it plays a central role in cellular energy regulation and serves as a sensitive marker of oxygen availability within tissues 42 , 43 . Lactic acid secretion is also associated with major tissues disfunction like lactic acidosis in fatty livers 44 . Lactate is also involved in the regulation of major physiological systems—including cardiovascular, respiratory, and digestive—and holds clinical significance in the diagnosis and prognosis of a range of conditions 45 . Given its biological importance and interpretability, lactic acid was selected as a demonstrator molecule and assess the feasibility of such online, non-destructive monitoring of metabolic activity within engineered tissues. In the field of cell culture bioprocessing, several analytical tools have been proposed for real-time metabolic monitoring. Among these, Raman spectroscopy shows great promise for non-invasive, continuous measurement of secreted or consumed metabolites 23 , 46 . In our study, we implemented a bIO-LAB 220 Raman probe directly within the perfusion circuit of the bioreactor vessel ( Fig. 2-A ), enabling the acquisition of spectra from culture supernatant circulating through the bioprinted engineered tissue ( Supplementary Figure 3 ). Unlike many conventional approaches, this setup allowed us to monitor metabolite concentrations without withdrawing samples, offering a real-time and label-free method of assessing tissue metabolic function. Our objective was to demonstrate the versatility of this method across different tissue types and cellular origins. To this end, engineered tissues were bioprinted using cell types from two distinct tissues from human origin, dermis fibroblasts and kidney epithelial cells (HEK). Importantly, lactate concentrations in tissue-based cultures are often in the range of 0.5-2 g.L -1 36 , compared to the 3–12 g.L -1 commonly encountered in suspension cultures 47 . This is a significant challenge for chemometrics model development, as standard Raman-based mathematical models are generally optimised for higher metabolite concentrations. We started our study by Raman calibration step. It consists in building a mathematical model correlating enzymatically quantified lactic acid concentrations with Raman spectral data. In order to access to Raman lactic acid signature within complex sample spectra, we had to use chemometric analysis. It involves building Orthogonal Partial Least Squares (OPLS) regression after spectra deconvolution in several principal components. This was performed thanks to Simca® 18 software (Sartorius, Germany). To build this calibration, we collected 76 discrete culture samples among 7 distinct tissue culture batches along culture which could last up to 32 days. To use best practice for Raman calibration 48 , we also apply sample spiking with added known amount of lactic acid in order to decorrelate its concentrations from other molecules present in the culture samples (ex. Glucose). Then, several spectral pre-processing methods were evaluated and screened —including baseline correction, normalisation, and first-order derivatives—to optimise the PLS model fit (R² CAL ), its predictive power (Q² CAL ), and to minimise root-mean-square errors (RMSE CAL ). Calibration step enabled to select the four best OPLS models ( Fig. 3-B ), from which Model 1 calibration curve is presented in Figure 3-A . To challenge the predictive performance of these four OPLS calibration models, an external dataset consisting of 21 additional Raman spectra, corresponding to quantified samples from four novel culture tissue batches were used to predict lactic acid. Prediction performance was evaluated by comparison between the real lactic acid quantified concentrations and the predicted concentration. The best predictive PLS model was the one combining spectral treatment with Asymmetric Least Squares (AsLS) baseline correction, Standard Normal Variate (SNV) normalisation, and 1 st derivative transformation. It enabled quantification of lactate concentrations within a range of 0 to 3.5 g.L -1 , with a prediction precision error RMSE PRED of 0.103 g.L -1 ( Fig. 3-B ) and a R 2 PRED of 0.942. Lactic acid Raman quantification was then deployed for the continuous online monitoring of 3D bioprinted conjunctive tissue composed of dermal fibroblasts perfused over a 13-day of culture. The Raman spectra of the perfused medium were acquired every 17 minutes from day 2 corresponding to a total of 933 spectra. The calculated lactate concentration evolution over time is presented in Figure 3-C . To facilitate lactic acid production trend readability, concentrations were averaged into hourly mean values and represented as black dots ( Fig. 3-C ). The predicted lactic acid production trend obtained is noisy as the production levels for dermis fibroblasts is very low, i.e. 0.2-0.3 g.L -1 , compared to the determined prediction performance of 0.1 g.L -1 . Still, lactic acid secretion trend is consistent with dynamics previously described in static culture for dermal fibroblasts 36 . Additionally at day 6, Raman probe was able to identify the lactic acid dilution implied by 50 mL medium additions, corresponding to 1/3 dilutions (see black arrow on Fig. 3-C ). This proof-of-concept illustrates that Raman spectroscopy can be successfully adapted for online metabolic monitoring in tissue engineering, despite the relatively low metabolite concentrations involved (0.2-0.3 g.L -1 ).These findings echo its established use in industrial-scale bioprocessing for molecules such as glucose, amino acids, and secreted proteins 23 , 45 . Although the development of a robust calibration model still requires a representative dataset of at least 75-100 samples—sufficient in both size and variability—the approach shows high potential for broader applications in engineered tissue monitoring. Going forward, the ability to monitor nutrient consumption and by-product accumulation in real time offers not only an advanced level of bioprocess control, but also a new window into tissue physiology and metabolic dynamics. Expanding this platform to include additional analytes and tissue types could pave the way for the creation of intelligent bioreactor systems capable of adapting in real time to the evolving needs of the engineered tissue. Morphology Monitoring of Engineered Tissues As engineered tissues mature in vitro , their internal structures undergo continuous remodelling driven by cellular activity. These changes include extracellular matrix (ECM) secretion and remodelling, formation of vessels and cavities, compaction, and overall densification of the tissue 49 , 50 . As a result, the geometry of 3D bioprinted engineered tissues is expected to evolve significantly throughout the culture period and must be monitored over time to ensure fidelity to the original design and to understand tissue development dynamics. To date, most techniques available for characterizing internal morphology are either destructive (histology or electron microscopy), require extensive sample preparation (light-sheet or confocal microscopy) 51 or lack depth of analysis (Optical coherence tomography) 28 , 52 . These constraints limit the ability to perform real-time, longitudinal monitoring of tissue morphology throughout maturation. Magnetic Resonance Imaging (MRI), widely recognized in vivo for its safety and non-invasiveness, offers a promising solution for the non-destructive monitoring of 3D engineered tissues 53 , 54 . In this study, we evaluated the application of MRI for tracking the morphological evolution of a 3D bioprinted engineered tissue seeded with human fibroblasts over a 15-day culture period. Figure 4 presents MRI data acquired using a 7 Tesla system with T2-weighted sequences, chosen for their high sensitivity to water content and thus their ability to distinguish hydrogels from surrounding fluids and matrix. High-resolution imaging was achieved without the use of any contrast agent, yielding isotropic voxels of 137×137×137 µm³. As shown in Figure 4-A , MR images acquired immediately after bioprinting offered clear visualization of the entire 10.8 cm³ engineered tissue, with excellent contrast between voids and matrix. Multiplanar views (sagittal, coronal, and axial) allowed comprehensive assessment of internal geometries, including nascent cavities. To quantitatively assess morphological fidelity, the segmented MRI data were compared with the original bioprinting STL design file ( Fig. 4-B ). This analysis, conducted using 3DSlicer (USA), compared over one million discrete spatial points across both models. Results showed that 87.6% of the MRI-derived tissue geometry deviated by less than 1 mm (representing ~4% of the engineered tissue's average dimension) from the original STL. These minor discrepancies are primarily attributed to extrusion-based bioprinting resolution limits and handling-induced deformations, rather than limitations of the imaging process itself 55 . To demonstrate longitudinal monitoring capabilities, the same engineered tissue was imaged after 16 days of culture under controlled physicochemical conditions ( Fig. 4-C ). MRI revealed significant morphological evolution, including pore occlusion (green star), shape deformation, and localized matrix degradation (red arrows). These features were corroborated by histological analysis of corresponding tissue sections (Masson’s trichrome staining, Fig. 4-D ), providing strong correlation between non-destructive and classical destructive approaches. Histology results also permit the identification of deposited extracellular matrix (green coloration) and embedded cells (black arrows), after 16 days of culture. Taken together, these results validate MRI as a reliable, non-destructive modality for high-resolution morphological monitoring of 3D bioprinted engineered tissues. The ability to acquire quantitative and longitudinal data without the need for labelling or sample preparation represents a major advancement in tissue engineering. This approach enables dynamic tracking of tissue maturation and deformation, contributing valuable insights for engineered tissue design optimization, quality control and regulatory documentation, all essential steps for advancing regenerative medicine applications. Perfusion Monitoring of Engineered Tissues As engineered tissues increase in volume and complexity, particularly beyond the cubic centimetre scale, simple diffusion mechanisms become insufficient to sustain cellular viability across the entire engineered tissue 31 . In such cases, active perfusion of culture media is required to ensure the homogeneous delivery of oxygen and nutrients while facilitating the removal of metabolic waste. This challenge is particularly relevant in the context of engineered tissues exceeding 10 cm³, such as those investigated in the present study. To design effective perfusion strategies and verify their implementation, we established a dual approach combining numerical simulation and experimental mapping of internal flow fields: computational fluid dynamics (CFD) was used to predict perfusion performance, and MRI velocimetry 56 was applied to non-destructively quantify fluid flow through the engineered tissues. We first employed CFD simulations using ANSYS Fluent to evaluate nutrient flow within the tissue perfusion vessel ( Fig. 2-A ) housing a 12×12 channels mock scaffold, 3D printed in PLA. Simulations were performed at two different flow rate, 2 mL.min -1 and 20 mL.min -1 . The resulting velocity maps, in the vertical plan including the inlet and outlet, are presented in Figure 4-A . A clearly heterogeneous distribution of the perfused liquid velocity was evidenced for both flow rates, consequence of the preferential horizontal flow path within the scaffold which presented lateral pores of 1.2x1.2 mm and vertical pores of 1.2x0.4 mm ( Supplementary Figure 1 ). Once these simulation performed, flow-sensitive MRI was used to experimentally measure internal fluid dynamics within the 12×12 channels mock scaffold subjected to identical flow conditions, i.e. 2 mL/min and 20 mL/min. In the particular case of this flow-sensitive MRI experiment, the previously used 7 Tesla MRI scanner was set up to achieve an isotropic resolution of 1 mm³. The system was tuned to detect flow velocities ranging from 200 µm.sec -1 to 10 cm.sec -1 , and imaging was performed without the use of any contrast agent. Figure 4-C present the velocity maps, in the vertical plan including the inlet and outlet, for the two tested flow rates. Strong agreement was observed in terms of global flow direction, velocity magnitude and pattern, when comparing MRI results to the CFD simulations. However, several discrepancies emerged mainly due to the presence of air bubbles in the experimental set-up. These bubbles, easily identified in the morphological image (white arrows in Fig. 4-B ) had a strong impact on the flow distribution. MRI-based velocimetry also enabled the visualisation of perfusion flows in 3D. Examples are given in Figure 4-D with 2 different views of the flow path volumes at the 2 different flow rates. Even though these representation are difficult to appreciate in 2D, their observation as 3D animation are of great interest to appreciate the internal flow geometries ( Supplementary Video 1 and 2 ). Finally, the complementary use of CFD and MRI provides a powerful framework for understanding and optimising perfusion in bioengineered tissues. CFD offers a rapid, flexible means of screening engineered tissue designs and bioreactor parameters in silico , whereas MRI yields spatially resolved, empirical data that can be acquired non-destructively and applied throughout the culture process. Importantly, MRI-based velocimetry is directly transferable to living engineered tissues and could be used to track changes in perfusion as a function of tissue maturation, cavity formation, or vascularisation. This combined approach lays the foundation for future development of feedback-regulated perfusion systems, capable of dynamically adjusting flow rates in response to tissue metabolic needs. Such systems would represent a significant advancement in the field of bioprocess control, enabling safer and more efficient maturation of large-scale engineered tissues for clinical and pharmaceutical applications. Discussion and Conclusion The increasing sophistication of biofabricated tissues, both in scale and biological complexity, demands a corresponding evolution in analytical strategies. As demonstrated in this study, non-destructive, real-time monitoring of engineered tissues is not only technically feasible but also essential for controlling the dynamic and multiscale processes underlying tissue maturation. We showed that the physicochemical environment, structural integrity, internal perfusion, and metabolic activity of 3D bioprinted engineered tissues can all be monitored using integrated sensor-based and imaging tools commonly available in pharmaceutical bioprocessing and clinical imaging, yet underexplored in tissue engineering contexts. The application of Raman spectroscopy for monitoring lactic acid secretion from engineered tissues constitutes a significant advance in metabolic surveillance. Through PLS modelling and spectrum pre-processing, we quantified lactate at physiologically relevant concentrations (from 0 to 3.5 g.L − 1 ) with a resolution of 0.10 g.L − 1 , enabling temporal mapping of metabolic activity every 17 minutes. This continuous readout correlates with known proliferation phases and provides a framework for applying metabolic control in future smart bioreactor platforms. Our work also highlights the critical importance of oxygen regulation in volumetric cultures. Standard cell culture protocols proved inadequate for large-scale engineered tissues, leading to significant deviations from target dissolved oxygen concentrations and reduced cell growth. By implementing dual-gas control and optimising PID parameters, we achieved a substantial improvement in oxygen regulation, reducing deviations from 131–11.5% in fibroblast culture. These improvements translated into better environmental stability and, likely, more reproducible biological outcomes. In terms of morphological monitoring , 7 Tesla MRI allowed for contrast agent-free, longitudinal tracking of 3D engineered tissues with a resolution sufficient to capture structural changes over time. MRI-derived geometries closely matched the original STL files (with over 87% of points deviating by less than 1 mm), and later scans revealed remodelling events such as pore closure and matrix compaction—validated by histology. This non-destructive approach offers valuable insights into tissue integrity and deformation throughout culture. Finally, our dual strategy for assessing tissue perfusion —via CFD simulation and MRI velocimetry—proved essential for validating flow distribution. Numerical models provided fast, scalable design tools, while MRI confirmed internal velocity profiles in printed engineered tissues, capturing flow heterogeneity as a function of architecture and flow rate. This capability supports the implementation of adaptive perfusion systems that can evolve with the tissue’s metabolic needs. Together, these tools outline a modular, scalable, and highly transferable analytical framework for engineered tissue monitoring. By adapting technologies already validated in pharmaceutical manufacturing, this study bridges a critical gap between traditional tissue culture and the emerging field of automated, closed-loop biofabrication. Beyond improving process control, these monitoring strategies lay the groundwork for regulatory alignment, standardisation, and quality assurance, both key steps toward the clinical translation of engineered tissues. Methods Cell culture Two cell types were chosen for this study: human dermal fibroblasts (obtained from Hospice Civil Lyon cell bank, France) and human kidney cells (HEK293T, provided by Dr. C. Maisse-Paradisi, INRA-UCBL-EPHE "Viral Infections and Comparative Pathology", Lyon, France). Fibroblasts and HEK293T were cultured in Dulbecco's Modified Eagle Medium (DMEM) (Thermo Fisher Scientific, 31966-021) supplemented with 10 % (v/v) FBS (Gibco Cell Culture, 10270-106). Cells were pre-cultured at 5 % CO 2 , 80 % humidity and 37°C before bioprinting. Fibroblasts and HEK293T were passaged once and twice a week, respectively. 3D Bioprinting of engineered tissues 3D bioprinting bioink was formulated using bovine gelatine (G1890 Sigma, France), very low viscosity alginate (A18565.36 Alfa Aesar, Thermo Fisher France) and fibrinogen from bovine plasma (F8630 Sigma, France). All components were handle under sterile laminar flow to ensure sterility. Stock solution of 0.2 g/mL gelatine, 0.04 g/mL alginate and 0.08 g/mL fibrinogen were dissolved, without any stirring for 18 hours at 37 °C, in DMEM (without calcium, with glutamax-1, Invitrogen, France) supplemented with 10% foetal calf serum (HyClone, USA), 20 µg/ml gentamicin (Pantapharm, France), 100 UI/ml penicillin/streptomycin (Sarbach, France) and 1 µg/ml amphotericin B (Bristol Myers Squibb, France). For bioink preparation, trypsinated cells were first suspended in calcium-free DMEM supplemented with 10 % FBS and enumerated. Targeted cell concentration was adjusted by pelleting the appropriate number of cells at 300 g (Fibroblasts, HEK293T) for 5 min. The pellet was suspended in the proper volume of 0.08 g/mL fibrinogen solution to reach 2.2 x 10 5 cells/mL for fibroblast and 2.6 x 10 6 cells/mL for HEK 293T. Then, to this cell suspension in fibrinogen, the appropriate volumes of alginate and gelatine stock solutions were added in order to reach a final composition of 0.02 mg/mL fibrinogen, 0.02 mg/mL alginate and 0.05 mg/mL gelatine. The bioink was homogenized and incubated for 15 min at 37°C. After homogenization, sterile cartridge (Nordson EFD, France) was filled with the bioink and incubated 30 minutes at 21°C to stabilise the bioink rheological properties. The cartridge was then loaded in a 6-axis robotic bioprinter (BioAssemblyBot®, Advanced Lifescience Solutions, USA) and used to print 10.8 cm 3 parallelepiped macroporous 3D structures (2x2x2.7 cm, 12x12 pores, internal pore dimensions of 1.2x1.2 mm). A 800 µm diameter, 6.35 mm long needle (Nordson EFD, France) was used to bioprint at a set speed an pressure of 8 mm/sec and 20-50 Psi, respectively. Once bioprinted, the tissues were consolidated with a solution composed of 40 mg/mL transglutaminase (TAG) (ACTIVA WM - Ajinomoto), 10 U/mL thrombin from bovine plasma (T4648-10KU - Merck) and 270 mM CaCl 2 (C5670-500G - Merck). The consolidation process was carried out at 37°C for 2 hours. The cellularised structures were then rinsed twice with sterile physiological serum (Versol, France) before being introduced in the tissue perfusion vessel. Engineered tissues perfusion chamber design and production Computer Assisted Design (CAD) of the tissue perfusion chamber was performed using Autodesk® Fusion 360™. CAD files, available in a GitHub repository (freely available at https://github.com/FabricAdvancedBiology/Multiflow_cell) were converted to .STL format to be transferred to an Object30 Pro inkjet printer (Stratasys, USA). The chamber was 3D-printed using VeroClear resin (Stratasys, USA). Once printed, the support material was removed using a high pressure waterjet (Stratasys). The 3D-printed chamber was then incubated overnight in 70 % ethanol at room temperature to remove all leachable compounds, and then finally rinsed in milliQ water. Prior any use for tissue perfusion, the chamber was steam sterilized at 120°C, 2 Bars for 20 minutes. Set up for physicochemical environment monitoring A controlled and regulated physicochemical tissue environment was established using a commercial bioreactor (Eppendorf DASbox® 250mL) connected to the 3D-printed tissue perfusion chamber through a peristaltic pumping systems. To monitor and control the physicochemical environment parameters (pH, dissolved oxygen, temperature), a custom Cytosys controller (Ipratech, Belgium) was employed. Customization of the system included an Applikon heating jacket 30W MINIBIO 250 heating blanket (Applikon, Sweden), a custom stirrer motor, 12 mm pH and oxygen probes (Hamilton) and two temperature probes (Pt100). The initial dissolved oxygen level regulation loop involved air and nitrogen injections, while pH was regulated thanks to CO 2 and 0.2 M NaOH solution. Before cultivation, Eppendorf DASbox® bioreactors was steam sterilised. A day before its connexion to the tissue perfusion chamber, the bioreactor was filled with culture medium and the pH, temperature and oxygen monitoring/regulation started to stabilise the probes’ response before culture. Once bioprinted and consolidated, the 10.8 cm 3 engineered tissues were sterilely inserted, under laminar flow, into the 3D-printed perfusion chamber. The tissue perfusion chamber was sealed, connected to perfusion tubing and the perfusion peristaltic pump started to generate perfusion at 2 or 20 mL/min perfusion rates. Lactic acid enzymatic assay Spent media were recovered daily through sampling, aliquoted in triplicate and stored at -20°C. These samples were further used to quantify the lactic acid produced by cell metabolism using the L-Lactic Acid Assay from Megazyme (L-Lactic Acid Assay Kit, K-LATE, Megazyme). The assay was performed in 96-well plates with an autosampler procedure according to the manufacturer's instructions. Each sample was analysed in triplicate at 340 nm with a spectrophotometer (TECAN infinite ® ). Samples presenting lactate concentration above the quantification linearity were diluted in deionised water. All results were normalised with a fresh culture medium used as a blank. Online Raman spectroscopy The acquisition of culture medium Raman spectra was performed online with an Rxn2 IOT Raman spectroscopy analyser (Endress Hauser) and associated Raman probe (bIO-LAB 220; Endress Hauser) kindly provided by Kaiser-Endress. The probe was placed within the DASbox® tank before set-up sterilisation. The excitation wavelength of the laser was 785 nm. A preliminary blank spectrum was acquired and lasted 15 minutes. The acquisition parameters were optimised to target between 50 to 80% of signal saturation. Thus, acquisition were set at 40 s with 20 accumulation counts thus corresponding to 17 minutes accumulation in total. After Raman spectrum acquisition, online lactate monitoring was calibrated based on offline samples. 76 off-line spectra were generated from spiked fresh or spent medium samples from fibroblast's culture with concentrated Lactic acid solution. Morphology monitoring of engineered tissues by Magnetic Resonance Imaging (MRI) At the end of the culture, engineered tissues were fixed by incubation in AntigenFix (DiaPath, P0014) overnight at 4°C. The engineered tissues were then rinsed and kept in ethanol 70 % before MRI analysis. The MRI morphology protocol was carried out on a 7 Tesla Bruker BioSpec MR system (Bruker Biospin GbmH, Germany) equipped with a 400 mT/m maximal amplitude gradient set and controlled using a Bruker workstation interfaced with ParaVision5.1 software for data acquisition and post-processing. A transmit-receive radio-frequency body coil (outer diameter 112 mm and inner diameter 72 mm) was employed for in vitro MR image acquisition. To acquire 3D high-resolution MR images of the engineered tissues, a tridimensional T2-weighted MR sequence based on the Rapid Acquisition with Relaxation Enhanced (RARE) method was performed on transversal orientation. The acquisition parameters were as follows; Echo Time (TE) 35.8 ms, Repetition Time (TR) 600 ms, bandwidth 75 kHz, RARE accelerator factor 8, and average number 2. A total of 256 slices of 273 µm thickness were acquired within a field of view of 3.5 x 3.5 x 3.5 cm 3 , and an in-plane matrix size of 256x256x128 interpolated to 256x256x256 pixels, providing a final in-plane isotropic resolution of 137x137x137 µm 3 . The total acquisition time was 1h1m26s. Perfusion Monitoring of Engineered Tissues by Magnetic Resonance Imaging Velocimetry Magnetic Resonance Imaging (MRI) protocol was carried out on a 7 Tesla Bruker BioSpec MR system. Mock scaffolds were 3D printed thanks to a Prusa MINI+ printer (PRUSA Research, Czech Republic) using 0.1 mm resolution slicing and a PLA filament (Prusament, PRUSA Research, Czech Republic). The structure of the mock scaffold is presented in Supplementary Figure 1 . They were sequentially inserted in the bioreactor’s chamber, leaving a gap of 2.2 mm towards the outlet wall and 2.4 mm towards 1 lateral wall. The bioreactor was then sealed and inserted inside the transmit-receive body coil positioned in the centre of the 7T MRI system ( Supplementary Figure 2 ). A flow loop was set up using a peristaltic pump, Masterflex Ismatec ISM834C. The whole set-up, including the rollers’ stretch, was connected with an autoclavable translucent hose, Versilic flexible tube, Ø 3mm x Ø 6mm (228-1194-VWR). The pump speed was calibrated by mapping the display reading to the displaced weighted volume, sweeping the whole range separated by 6 sampled speeds. A linear regression model was established (coefficient of determination R 2 = 0.993). The hydraulic circuit was filled with stained water. The incoming flow hose was connected to one bottom orifice, while the outgoing flow hose was released in the opposite-side upper orifice. The pump was set to run the experiments with two different pump flows: 2 mL/min and 19.6 mL/min. The velocity map sequence acquisition was carried out using the flow-imaging technique known as Flow Map, which is based on the phase contrast method 22 . Phase contrast MRI (PC-MRI) is a non-invasive imaging technique that does not require contrast agents and can provide high-resolution images of blood flow in any direction. PC MRI relies on the principle that the phase of the MRI signal is proportional to the velocity of the protons. To perform PC-MRI, two images are acquired with oppositely applied magnetic field gradients. This creates a phase difference between the two images that is proportional to the velocity of the protons. The phase difference image can then be used to calculate the velocity of the flowing fluid. The sequence was set with the following parameters: average of four images, eco time: 5 ms, repetition time: 50 ms. The maximum flow velocity measured with a given flow imaging sequence was set at 20 cm.s -1 . In comparison, the minimum flow velocity measured with the given pulse sequence parameters was set at 0.74 cm.s -1 . Four scenarios, two velocities for two macroporous structures, were acquired sweeping consecutive sagittal planes (XY) separated by 1mm and encoding the velocity vector components in all directions (XYZ). Each acquisition was set to cover a field of view of 5 x 5 cm, represented in 256 x 256 pixel images with a resolution of 195 μm.pixel -1 for each direction (XYZ). The photos were encapsulated both in DICOM and 2D sequence formats. Each experiment was completed in 58 min. Morphological images were also acquired in the same sweeping positions applying the T2-weighted spin-echo sequence, with and without running flow. Post-processing treatment of velocimetry data The raw data captured in the binary file format owned by BRUKER, namely, 2D sequence data (.2dseqs) and then it was decoded, using the BRUKER’s Application Programming Interface (API) installed in the programming language Python and converted from a four-dimensional array to segregated images in text file format. The 92 images corresponding to the 23 cuts were loaded in a MATLAB to be post-processed. The Euclidean norm of the vector velocity was calculated by combining the vector components in the three directions and scaling the result to obtain a value in meters per second. All the images were inspected to spot the bubbles' cross-sections, characterized by a region of high heterogeneity of color (spatial noise). Each image was treated individually to take the values in the bubble region to 0 m.sec -1 . The morphological images were used to create a binary mask to remove the noise from the periphery of the liquid chamber. The 3-dimensional representation of the velocity norm in all the volume was obtained by upsampling the images in the Z direction using the interpolation protocol defined in [10.1038/srep29936] to even out the resolution to 195 μm/slide and keep the real proportions. The new set of 110 images was stacked and rendered in the software 3D slicer using a smoked appearance to maximize the overall view in one shot. Histological analyses Fixed engineered tissues, stored in 70% ethanol were paraffin embedded using successive incubations at 65°C in 95% ethanol, 100% ethanol, ethanol/butanol mix, and butanol/methylcyclohexane mixes. Each incubation lasted for 20 min. Paraffin impregnation was then performed at 56°C for 2 h. Paraffin blocks were formed using metal moulds. The dehydrated engineered tissues were sliced into thin sections of 5 µm using a microtome. The sections were disposed of on Superfrost slides (Epredia, 12321). The sections were stained for Masson’s Trichrome to observe cell and collagen repartition inside the structure. Sections were stained with Groat’s hematoxylin for 1 min, Fuschin-ponceau for 3 min, G orange-phosphomolybdic acid for 3 min, and fast green for 3 min. Each staining step was followed by a water or 1% acetic acid wash. Sections were then dehydrated by rapid immersions in 70%, 95%, and 100% ethanol, and finally, slides were mounted with ExPert Mounting Medium (CellPath, SEA-1604-00A). Computational fluid dynamics (CFD) Fluid dynamics simulations were performed using the commercial finite volume-solver ANSYS Fluent2019. Based on the previous work of Pourchet et al. 23 , a single-phase with laminar flow was supposed, and a steady-state approach was applied for the solving of both continuity and momentum Navier-Stokes equations (Eq. (1) and (2)), with ⍴ the fluid density, v the fluid velocity, p the pressure, 𝜏 the Reynolds stress tensor and g the acceleration due to gravity. Fluid properties were like those of water at 37°C, namely ρ = 993 kg m -3 and dynamic viscosity µ = 6.92 × 10 -4 Pa.s. Parameters discretisation was achieved by a mesh divided into approximately 19 million tetrahedral meshes. The perfusion was performed by setting a fixed flow rate and pressure outlet corresponding to the flow inlet/outlet. All other boundaries were set as no-slip walls. Finally, solving equations used the SIMPLE method for the pressure-velocity coupling and 2 nd- order UPWIND schemes for pressure and momentum transport equations. Simulations were run until both a minimum criterion on residuals (< 10 -3 ) and stabilisation of the fluid velocity, the pressure outlet and the wall shear stress were reached. CFD post-treatment was performed using ANSYS CFD-Post software. According to the tissue internal flow monitoring, specific velocity flow fields were plotted to compare the experimental data with the modelled ones. Declarations Funding Declaration The authors declare no funding relate to the present study. Author Contribution LC, SP, NEK, EP, SL, and CL realised the experiments and generated the data. CM, YG, SL, KM, CL, and EP performed data curing, analysis, model building, and figure building. EP, CM, YG, and CL participated in the manuscript writing. All authors reviewed the manuscript. Acknowledgement We express our sincere gratitude to Jean Lynce Gnanago for his pivotal role in initiating the research on MRI flow analysis during his PhD. His dedicated study forms the cornerstone of the work presented in this paper. We acknowledge Radu Bolbos and CERMEP MRI 7T for their invaluable support in facilitating the MRI acquisition for this study—special thanks to the Sartorius PAT Team for their crucial assistance in Raman's chemometric analysis. We also acknowledge Sartorius for their financial support, which contributed to completing this work, and for providing the SIMCA software. We also acknowledge Kaiser Endress-Hauser and, in particular, Thomas Perilli for the loan of Raman equipment Rxn2 and the associated probes. The collaborative efforts and support from these individuals and organisations have been instrumental in the successful execution and completion of this research project. Data Availability Data will be available upon request to the corresponding author. References Pourchet, L. et al. Large 3D bioprinted tissue: Heterogeneous perfusion and vascularization. Bioprinting 13 , (2019). Marquette, C. A. et al. Unlocking the potential of bio-inspired bioinks: A collective breakthrough in mammalian tissue bioprinting. Bioprinting 41 , (2024). Li, H. et al. Digital light processing (DLP)-based (bio)printing strategies for tissue modeling and regeneration. Aggregate 4 , (2023). Li, W. et al. Stereolithography apparatus and digital light processing-based 3D bioprinting for tissue fabrication. iScience 26 , (2023). Shiwarski, D. J. et al. Emergence of FRESH 3D printing as a platform for advanced tissue biofabrication. APL Bioeng 5 , (2021). Bernal, P. N. et al. Volumetric Bioprinting of Complex Living-Tissue Constructs within Seconds. Adv Mater 31 , (2019). Zhang, Y. S. et al. 3D extrusion bioprinting. Nature Reviews Methods Primers 1 , (2021). Duarte, C. D. F. et al. Synchronized Dual Bioprinting of Bioinks and Biomaterial Inks as a Translational Strategy for Cartilage Tissue Engineering. 3d Print Addit Manuf 6 , (2019). Ali, A. S. M. et al. 3D bioprinting of liver models: A systematic scoping review of methods, bioinks, and reporting quality. Materials Today Bio 26 , (2024). Fang, Y. et al. 3D printing of vascularized hepatic tissues with a high cell density and heterogeneous microenvironment. Biofabrication , (2023). Grigoryan, B. et al. Multivascular networks and functional intravascular topologies within biocompatible hydrogels. Science 364 , (2019). Lee, A. et al. 3D bioprinting of collagen to rebuild components of the human heart. Science 365 , (2019). Skylar-Scott, M. A. et al. Biomanufacturing of organ-specific tissues with high cellular density and embedded vascular channels. Sci Adv 5 , (2019). Lee, S. J. et al. 3D Bioprinting for Engineered Tissue Constructs and Patient-Specific Models: Current Progress and Prospects in Clinical Applications. Adv Mater 36 , (2024). Edri, S. et al. 3D Bioprinting of Multicellular Stem Cell-Derived Constructs to Model Pancreatic Cell Differentiation. Adv Func Mater 34 , (2024). Scognamiglio, C. et al. Bioprinting stem cells: building physiological tissues one cell at a time. American Journal of Physiology-Cell Physiology 319 , (2020). Leferink, A. M. et al. Methods of Monitoring Cell Fate and Tissue Growth in Three-Dimensional Scaffold-Based Strategies for In Vitro Tissue Engineering. Tissue Engineering Part B: Reviews 22 , (2016). Mather, M. L. et al. Meeting the Needs of Monitoring in Tissue Engineering. Regenerative Medicine 2 , (2007). Holzberg, T. R. et al. Sensors for biomanufacturing process development: facilitating the shift from batch to continuous manufacturing. Current Opinion in Chemical Engineering 22 , (2018). Visone, R. et al. Enhancing all-in-one bioreactors by combining interstitial perfusion, electrical stimulation, on-line monitoring and testing within a single chamber for cardiac constructs. Scientific Reports 8 , (2018). Zoneff, E. et al. Controlled oxygen delivery to power tissue regeneration. Nature communications 15 , (2024). Eshmuminov, D. et al. An integrated perfusion machine preserves injured human livers for 1 week. Nat Biotechnol 38 , (2020). Whelan, J. et al. In situ Raman spectroscopy for simultaneous monitoring of multiple process parameters in mammalian cell culture bioreactors. Biotechnol Prog 28 , (2012). Stepula, E. et al. Label-free 3D molecular imaging of living tissues using Raman spectral projection tomography. Nature communications 15 , (2024). Clegg, J. et al. Tailoring metabolic activity assays for tumour-engineered 3D models. Biomaterials Advances 167 , (2025). Martin, A. B. et al. Human Anatomy and Physiology I . (LOUIS: The Louisiana Library Network, 2024). Kotecha, M. in Magnetic Resonance Imaging in Tissue Engineering 21-47 (2017). Gil, C. J. et al. Leveraging 3D Bioprinting and Photon-Counting Computed Tomography to Enable Noninvasive Quantitative Tracking of Multifunctional Tissue Engineered Constructs. Adv Healthc Mater 12 , (2023). Liang, X. et al. Imaging engineered tissues using structural and functional optical coherence tomography. Journal of biophotonics 2 , (2009). Tan, W. et al. Optical coherence tomography of cell dynamics in three-dimensional tissue models. Optics Express 14 , (2006). Bland, E. et al. Overcoming hypoxia to improve tissue-engineering approaches to regenerative medicine. J Tissue Eng Regen M 7 , (2013). Ehrbar, M. et al. Editorial: Advancing vascularized tissue models through biomaterials and biofabrication. Front Bioeng Biotechnol 12 , (2024). Place, T. L. et al. Limitations of oxygen delivery to cells in culture: An underappreciated problem in basic and translational research. Free Radical Biology and Medicine 113 , (2017). Wagner, B. A. et al. The rate of oxygen utilization by cells. Free Radic Biol Med 51 , (2011). Magliaro, C. et al. Oxygen Consumption Characteristics in 3D Constructs Depend on Cell Density. Frontiers in Bioengineering and Biotechnology Volume 7 - 2019 , (2019). Chastagnier, L. et al. Deciphering dermal fibroblast behavior in 3D bioprinted dermis constructs. Bioprinting 32 , (2023). Harcum, S. W. et al. PID controls: the forgotten bioprocess parameters. Discover Chemical Engineering 2 , (2022). Jones, J. et al. Improved control strategies for the environment within cell culture bioreactors. Food and Bioproducts Processing 138 , (2023). Chen, Y. et al. Characterization and optimization of cell seeding in scaffolds by factorial design: quality by design approach for skeletal tissue engineering. Tissue Eng Part C Methods 17 , (2011). Martinez-Marquez, D. et al. Application of quality by design for 3D printed bone prostheses and scaffolds. PLOS ONE 13 , (2018). Papantoniou, I. et al. Process Quality Engineering for Bioreactor-Driven Manufacturing of Tissue-Engineered Constructs for Bone Regeneration. Tissue Engineering Part C: Methods 19 , (2012). Boubriak, O. A. et al. Monitoring of metabolite gradients in tissue-engineered constructs. Journal of the Royal Society, Interface 3 , (2006). Lagarde, D. et al. Lactate Fluxes and Plasticity of Adipose Tissues: A Redox Perspective. Frontiers in Physiology Volume 12 - 2021 , (2021). Yao, S. et al. Role of lactate and lactate metabolism in liver diseases (Review). Int J Mol Med 54 , (2024). Li, X. et al. Lactate metabolism in human health and disease. Signal Transduction and Targeted Therapy 7 , (2022). Bhatia, H. et al. In-line monitoring of amino acids in mammalian cell cultures using raman spectroscopy and multivariate chemometrics models. Eng Life Sci 18 , (2017). Esmonde-White, K. A. et al. The role of Raman spectroscopy in biopharmaceuticals from development to manufacturing. Anal Bioanal Chem 414 , (2022). Ntziouni, A. et al. Review of Existing Standards, Guides, and Practices for Raman Spectroscopy. Appl Spectrosc 76 , (2022). Urciuolo, F. et al. In vitro strategies for mimicking dynamic cell–ECM reciprocity in 3D culture models. Frontiers in Bioengineering and Biotechnology Volume 11 - 2023 , (2023). Yi, J.-H. et al. Remodeling of soft tissues due to cell activity. PAMM 14 , (2014). Andilla, J. et al. Imaging tissue-mimic with light sheet microscopy: A comparative guideline. Scientific Reports 7 , (2017). Matcher, S. J. Practical aspects of OCT imaging in tissue engineering. Methods Mol Biol 695 , (2011). Hahn, A. et al. in Tissue Engineering Using Ceramics and Polymers (Third Edition) (eds Aldo R. Boccaccini, Peter X. Ma, & Liliana Liverani) 281-343 (Woodhead Publishing, 2022). Xu, H. et al. Monitoring tissue engineering using magnetic resonance imaging. J Biosci Bioeng 106 , (2008). Guida, L. et al. Advancements in high-resolution 3D bioprinting: Exploring technological trends, bioinks and achieved resolutions. Bioprinting 44 , (2024). Ricke, A. et al. Magnetic Resonance Velocimetry for porous media: sources and reduction of measurement errors for improved accuracy. Experiments in Fluids 64 , (2023). Additional Declarations No competing interests reported. Supplementary Files Chastagnieretal.Supplementary.docx Cite Share Download PDF Status: Published Journal Publication published 07 Jan, 2026 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Revision requested 08 Oct, 2025 Reviewers agreed at journal 18 Sep, 2025 Reviews received at journal 15 Sep, 2025 Reviewers agreed at journal 14 Sep, 2025 Reviews received at journal 07 Sep, 2025 Reviewers agreed at journal 17 Aug, 2025 Reviewers invited by journal 14 Aug, 2025 Editor assigned by journal 14 Aug, 2025 Editor invited by journal 12 Aug, 2025 Submission checks completed at journal 08 Aug, 2025 First submitted to journal 08 Aug, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7216794","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":501551475,"identity":"ddc116e2-a7a4-4663-8eec-427d7efff3eb","order_by":0,"name":"Laura Chastagnier","email":"","orcid":"","institution":"3d.FAB, CNRS, INSA, CPE-Lyon, Université Claude Bernard Lyon 1","correspondingAuthor":false,"prefix":"","firstName":"Laura","middleName":"","lastName":"Chastagnier","suffix":""},{"id":501551476,"identity":"4f5d0460-64cc-4b66-b50c-24dc11e38837","order_by":1,"name":"Sarah Pragnere","email":"","orcid":"","institution":"3d.FAB, CNRS, INSA, CPE-Lyon, Université Claude Bernard Lyon 1","correspondingAuthor":false,"prefix":"","firstName":"Sarah","middleName":"","lastName":"Pragnere","suffix":""},{"id":501551477,"identity":"598b04e7-8c3a-4bb5-bcf5-97c7cecbbe3e","order_by":2,"name":"Yilbert Gimenez","email":"","orcid":"","institution":"3d.FAB, CNRS, INSA, CPE-Lyon, Université Claude Bernard Lyon 1","correspondingAuthor":false,"prefix":"","firstName":"Yilbert","middleName":"","lastName":"Gimenez","suffix":""},{"id":501551480,"identity":"c8a06765-72e3-4758-beae-680a43eb9321","order_by":3,"name":"Celine Loubière","email":"","orcid":"","institution":"Université de Lorraine, CNRS, LRGP","correspondingAuthor":false,"prefix":"","firstName":"Celine","middleName":"","lastName":"Loubière","suffix":""},{"id":501551481,"identity":"180b128a-d4e7-4eaa-baff-f7e455f24e0b","order_by":4,"name":"Lucie Essayan","email":"","orcid":"","institution":"3d.FAB, CNRS, INSA, CPE-Lyon, Université Claude Bernard Lyon 1","correspondingAuthor":false,"prefix":"","firstName":"Lucie","middleName":"","lastName":"Essayan","suffix":""},{"id":501551482,"identity":"226cd395-224f-4795-b778-6dbd1468c67d","order_by":5,"name":"Kleanthis Mazarakis","email":"","orcid":"","institution":"Sartorius Stedim UK Ltd","correspondingAuthor":false,"prefix":"","firstName":"Kleanthis","middleName":"","lastName":"Mazarakis","suffix":""},{"id":501551483,"identity":"bd302170-fb4f-4bcf-9fea-bf5b5b5bd634","order_by":6,"name":"Timo Schmidberger","email":"","orcid":"","institution":"Sartorius Stedim Biotech GmbH","correspondingAuthor":false,"prefix":"","firstName":"Timo","middleName":"","lastName":"Schmidberger","suffix":""},{"id":501551484,"identity":"5a1a05d0-15bc-446b-a962-251159d24c95","order_by":7,"name":"Eric Olmos","email":"","orcid":"","institution":"Univ Lyon, Université Claude Bernard Lyon 1, INSA Lyon, Ecole Centrale de Lyon, CNRS, Ampère, UMR5005","correspondingAuthor":false,"prefix":"","firstName":"Eric","middleName":"","lastName":"Olmos","suffix":""},{"id":501551485,"identity":"618057f3-049c-4cd8-a3e5-e37102216eca","order_by":8,"name":"Simon Auguste Lambert","email":"","orcid":"","institution":"Ecole Centrale Lyon","correspondingAuthor":false,"prefix":"","firstName":"Simon","middleName":"Auguste","lastName":"Lambert","suffix":""},{"id":501551486,"identity":"12b7ad8e-6061-4059-b9aa-d4dab45b2551","order_by":9,"name":"Christophe A. Marquette","email":"","orcid":"","institution":"3d.FAB, CNRS, INSA, CPE-Lyon, Université Claude Bernard Lyon 1","correspondingAuthor":false,"prefix":"","firstName":"Christophe","middleName":"A.","lastName":"Marquette","suffix":""},{"id":501551487,"identity":"f3a9075c-3372-4437-b81a-9ee3f4eabe2c","order_by":10,"name":"Emma Petiot","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABC0lEQVRIie3RsWoCMRjA8S8Ech1ib5PvsNRXSAmog+izHEK3gOBYOE4K6di1UHwHRZAbhQOX3gMUbrE4F65bh0LNpZPQeGuh+U/5ID+SEACf7y/HAMgeAAGCehQQNglkQKmwhNbzFKK0iZitDO3SkspNwvZLXJFsmFwGj7u7VjaIU8pvDlMB2HeY6FmtkRS3yHjOylaBNZHyyZCr7e9ElGqJROfIcGKIRjnOea/DBSTouNi4VOtPor8tmdUE6A9BFxEdtTGnbC2hhlw3EizVZhAXk0jzXEYLS9hMcoFOEprnv35ko7D7MH+r3nXCIbhfHfjX0ElscXoyXgiwH3UuckqC/fntPp/P9986AsLEQ+P2O2RMAAAAAElFTkSuQmCC","orcid":"","institution":"3d.FAB, CNRS, INSA, CPE-Lyon, Université Claude Bernard Lyon 1","correspondingAuthor":true,"prefix":"","firstName":"Emma","middleName":"","lastName":"Petiot","suffix":""}],"badges":[],"createdAt":"2025-07-25 19:08:07","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7216794/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7216794/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-025-29257-y","type":"published","date":"2026-01-07T15:58:29+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":89664724,"identity":"d62e9590-936b-4de4-bc83-bf69e7af7ddd","added_by":"auto","created_at":"2025-08-22 11:51:45","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":382104,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eGeneral approach of the multiple monitoring modalities for large-scale 3D bioprinted tissue cultivation.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7216794/v1/6836b8cc84a266f7b2f6fd08.png"},{"id":89665395,"identity":"def2b54d-140c-4eda-98a5-1e98386cdb3a","added_by":"auto","created_at":"2025-08-22 11:59:45","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":367687,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eOnline real-time regulation of tissue physicochemical environment\u003c/strong\u003e. \u003cstrong\u003eA. \u003c/strong\u003eScheme of the experimental set up for the perfusion culture of bioprinted engineered tissues. \u003cstrong\u003eB.\u003c/strong\u003e Bioprinted engineered tissue perfusion culture monitoring using initial protocol showing discrepancy between set point and measured values. \u003cstrong\u003eC.\u003c/strong\u003e Bioprinted engineered tissue perfusion culture monitoring using optimized protocol and PID values (\u003cem\u003e\u003cstrong\u003eSupplementary Material Table 1)\u003c/strong\u003e\u003c/em\u003e. \u003cstrong\u003eD.\u003c/strong\u003e Bioprinted engineered tissue perfusion culture monitoring using optimized protocol, PID values and set points. Arrows identify deviations introduced due to manipulation (media addition in the regulated vessels, sampling, etc…). dO\u003csub\u003e2\u003c/sub\u003e represents the dissolved oxygen measured in % of air saturation.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7216794/v1/14d26c44884bafbc79816844.png"},{"id":89664526,"identity":"6f763f16-58db-4e05-abe5-bed0fba1c56e","added_by":"auto","created_at":"2025-08-22 11:43:45","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":224372,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eNon-destructive online monitoring of lactic acid secretion. A. \u003c/strong\u003eExample of calibration of Raman lactic acid quantification thanks to correlation with measured lactic acid concentration. Data presented correspond to OPLS Model 1 built from cultures performed with human dermis fibroblasts or kidney epithelial cells (HEK). \u003cstrong\u003eB.\u003c/strong\u003e Table presenting four OPLS calibration model performances for lactic acid quantification. \u003cstrong\u003eC.\u003c/strong\u003e Real-time and online monitoring of lactic acid production from bioprinted human dermis fibroblasts, using Raman spectra acquisition and Model 1 data treatment.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7216794/v1/deb5e0108e89f5eac310a7f2.png"},{"id":89664532,"identity":"edd00daf-2229-4bd0-b576-94e37d9f6c16","added_by":"auto","created_at":"2025-08-22 11:43:45","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":2111688,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMagnetic Resonance Imaging analysis of a 10.8 cm\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003e3\u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003e 3D bioprinted engineered tissue seeded with human fibroblasts.\u003c/strong\u003e \u003cstrong\u003eA.\u003c/strong\u003e Optical and MR imaging of the tissue at Day 0. \u003cstrong\u003eB. \u003c/strong\u003eExample of a numerical analysis of the tissue geometrical characterisation using volume comparisons (Artec Studio 18 (Artec 3D, Luxemburg)). \u003cstrong\u003eC.\u003c/strong\u003e MR image analysis of the same tissue after 16 days of culture in the regulated bioreactor (\u003cstrong\u003eFig. 2-A\u003c/strong\u003e). Axial and Sagittal views. \u003cstrong\u003eD.\u003c/strong\u003e Corresponding histological section (Masson’s trichrome staining). Green stars indicate pore occlusion, red arrows indicate localized matrix degradation, and black arrows indicate cells within the gel matrix.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7216794/v1/07dfc1c2652d9b89484bdd9f.png"},{"id":89664535,"identity":"893a8fac-7837-441c-a762-662ae0438df0","added_by":"auto","created_at":"2025-08-22 11:43:45","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1155777,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eNon-destructive characterisation of flow distribution within a 12x12 pores PLA 3D architecture by 7 Tesla MRI. A. \u003c/strong\u003eNumerical simulation of nutritive flow paths within 3D architectures. \u003cstrong\u003eB.\u003c/strong\u003e Imaging of internal morphology showing the presence air bubble (white arrows) and a gap between the scaffold and the vessel wall (red arrow). Red rectangular define the ROI for flow analysis. \u003cstrong\u003eC.\u003c/strong\u003e Planar velocity map in the same plan as A. \u003cstrong\u003eD.\u003c/strong\u003e 3D views of the flow volumes after segmentation (thresholds: 0-3-9 mm.sec\u003csup\u003e-1\u003c/sup\u003e).\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7216794/v1/aedeb474e8946c863461d062.png"},{"id":100070385,"identity":"06d130d6-a1da-4695-8b69-890e08ac061e","added_by":"auto","created_at":"2026-01-12 16:17:37","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":5664828,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7216794/v1/14cd7069-e868-42a1-b98b-d04d2583931b.pdf"},{"id":89665396,"identity":"baf5c3c1-8635-4b9c-ab60-27193a4a4660","added_by":"auto","created_at":"2025-08-22 11:59:45","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":1818836,"visible":true,"origin":"","legend":"","description":"","filename":"Chastagnieretal.Supplementary.docx","url":"https://assets-eu.researchsquare.com/files/rs-7216794/v1/6ad59a1215cb596ee7b88159.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Exploring multiple monitoring modalities for large-scale 3D tissue cultivation in bioreactor","fulltext":[{"header":"Introduction","content":"\u003cp\u003eThe field of engineered tissue biofabrication is undergoing a transformative evolution, rapidly progressing toward the fabrication of fully functional, implantable human organs. This progress is driven by the convergence of advances in 3D bioprinting, stem cell engineering, biomaterials science, and bioinformatics. Today, several complementary techniques are being developed to reconstruct tissue architecture and function, including extrusion-based bioprinting\u003csup\u003e1\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e2\u003c/sup\u003e, digital light processing (DLP)\u003csup\u003e3\u003c/sup\u003e, stereolithography\u003csup\u003e4\u003c/sup\u003e, freeform reversible embedding of suspended hydrogels (FRESH)\u003csup\u003e5\u003c/sup\u003e, and volumetric bioprinting\u003csup\u003e6\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eExtrusion-based bioprinting remains the most widely used due to its versatility and compatibility with a broad range of bioinks\u003csup\u003e7\u003c/sup\u003e. It has been applied to fabricate cartilage\u003csup\u003e8\u003c/sup\u003e, liver-like tissues\u003csup\u003e9\u003c/sup\u003e, and vascular networks with high precision\u003csup\u003e10\u003c/sup\u003e. More recently, volumetric bioprinting has emerged as a breakthrough technology, enabling the rapid fabrication of centimeter-scale soft tissues within seconds through tomographic light projections, achieving high spatial fidelity and cellular viability \u003csup\u003e11\u003c/sup\u003e\u003c/p\u003e\n\u003cp\u003eThe FRESH technique\u0026mdash;where bioinks are printed into a gelatin microparticle support bath\u0026mdash;has enabled the successful printing of complex, soft, and delicate structures such as collagen-based heart valves and mini hearts, with preserved geometry and mechanical properties\u003csup\u003e12\u003c/sup\u003e. Additionally, researchers have used DLP-based bioprinters to construct multiscale vasculature capable of supporting perfusable organoids, a significant step toward the vascularization of thick tissues\u003csup\u003e13\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eThese advancements are further supported by the integration of patient-specific imaging data and machine learning algorithms that allow personalized of biofabricated engineered tissues\u003csup\u003e14\u003c/sup\u003e. Coupled with the increasing maturity of stem cell-derived organoids and tissue-specific progenitor cells\u003csup\u003e15\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e16\u003c/sup\u003e, these technologies collectively bring the field closer to the production of clinically relevant, transplantable organs.\u003c/p\u003e\n\u003cp\u003eHowever, as the complexity and dynamism of engineered tissues grow, traditional analytical methods are reaching their limitations. Most current approaches rely on destructive assays or endpoint analyses, which are incompatible with dynamic, iterative, and long-term biomanufacturing processes\u003csup\u003e17\u003c/sup\u003e. To ensure quality and functionality throughout tissue maturation, the field requires an integrated, non-invasive monitoring environment capable of providing real-time feedback during and after fabrication\u003csup\u003e18\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eTo facilitate this paradigm shift, a tailored analytical framework must be developed to support next-generation biofabrication workflows. This framework should be capable of capturing and interpreting the most critical parameters that define the maturation and performance of engineered tissues. Among these, four classes of parameters are of particular importance: physicochemical environment, metabolic activity, morphology and perfusion.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e1. Physicochemical Environment. \u003c/strong\u003eThe local tissue environment\u0026mdash;including pH, oxygen tension, osmolarity, and temperature\u0026mdash;is a key regulator of cell behavior, influencing proliferation, differentiation, and apoptosis. Deviation from optimal conditions can impair tissue function and lead to cellular stress. Embeddable sensors capable of real-time, non-invasive tracking of these physicochemical parameters, including optical and electrochemical sensors for pH and oxygen, have been successfully integrated into bioreactor\u003csup\u003e19\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e20\u003c/sup\u003e. These tools allow the reconstruction of physiological gradients essential for mimicking \u003cem\u003ein vivo\u003c/em\u003e-like tissue environments\u003csup\u003e21\u003c/sup\u003e, such as metabolic zonation in liver engineered tissues.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2. Metabolic Activity. \u003c/strong\u003eMetabolic readouts provide a functional assessment of cellular health and activity\u003csup\u003e22\u003c/sup\u003e. Parameters such as glucose uptake, lactate production, or mitochondrial respiration reflect the bioenergetic state of cells within the engineered tissue. These metrics are particularly useful for evaluating differentiation status or the response to environmental changes. For instance, Raman spectroscopy has been used to monitor metabolic shifts \u003csup\u003e23\u003c/sup\u003eassociated with stem cell differentiation without the need for destructive sampling\u003csup\u003e24\u003c/sup\u003e. Real-time metabolic monitoring can guide media exchange, differentiation protocols, and overall culture optimization\u003csup\u003e25\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3. Morphology. \u003c/strong\u003eTissue morphology, encompassing both micro- and macro-architectural organization, is foundational to the structural integrity and functional identity of engineered tissues\u003csup\u003e26\u003c/sup\u003e. Proper spatial arrangement of cells and extracellular matrix is vital for enabling tissue-specific functions such as nutrient diffusion, signal propagation, and mechanical resistance. Non-invasive imaging techniques like magnetic resonance imaging (MRI)\u003csup\u003e27\u003c/sup\u003e and optical coherence tomography (OCT)\u003csup\u003e28\u003c/sup\u003e have been successfully applied to monitor morphology over time in 3D scaffolds and hydrogels\u003csup\u003e29\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e30\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4. Perfusion. \u003c/strong\u003eAdequate perfusion is essential to supply cells with oxygen and nutrients while removing metabolic waste products. In thick or vascularized engineered tissues, insufficient perfusion can lead to hypoxic zones and cell death, compromising the viability and function of the entire tissue\u003csup\u003e31\u003c/sup\u003e. Monitoring perfusion helps assess the effectiveness of biofabricated vasculature or perfusion bioreactors and is critical in scaling engineered tissues toward clinical dimensions\u003csup\u003e32\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cbr\u003eTo address the pressing need for real-time, non-invasive monitoring of biofabricated engineered tissues, the present study investigates a suite of integrated, non-destructive tools designed to assess critical parameters during tissue development (\u003cstrong\u003eFig. 1\u003c/strong\u003e). Specifically, we explore the monitoring and qualification of engineered tissues with a focus on the control of physicochemical conditions, compositional dynamics, and structural evolution over time.\u003c/p\u003e\n\u003cp\u003eOur approach leverages a combination of advanced sensing and imaging technologies\u0026mdash;including embedded temperature, pH and oxygen sensors, magnetic resonance imaging (MRI), and vibrational spectroscopy (e.g., Raman analysis)\u0026mdash;to assess their respective and complementary contributions to a robust analytical framework. These tools are evaluated not only for their technical feasibility within a biofabrication context but also for their ability to support dynamic, adaptive tissue engineering workflows.\u003c/p\u003e\n\u003cp\u003eBy integrating these modalities into the biofabrication pipeline, we aim to demonstrate the feasibility of establishing a real-time, feedback-driven system for tissue quality assessment\u0026mdash;an essential component for future scalable and clinically relevant tissue regeneration strategies.\u003c/p\u003e\n"},{"header":"RESULTS","content":"\u003cp\u003e\u003cstrong\u003ePhysicochemical Environment Monitoring in Engineered Tissues\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTissue maturation protocols are typically developed on small-scale engineered tissues, often limited to the millimetre-cube (mm\u0026sup3;) range, within partially controlled physicochemical environments. These studies are generally conducted in incubators with regulated temperature and buffered pH, yet oxygen delivery remains unregulated\u0026mdash;despite its pivotal role in cell physiology. Such limited control becomes a significant bottleneck when attempting to scale protocols to larger engineered tissues in the centimetre-cube (cm\u0026sup3;) range.\u003c/p\u003e\n\u003cp\u003eOxygen, in particular, plays a critical role in tissue development and regeneration, as reviewed by Zoneff et al. \u003csup\u003e21\u003c/sup\u003e. During the proliferative phase, oxygen is required for glycoprotein biosynthesis, while in the maturation and remodelling phases, it acts as a cofactor in hydroxyproline synthesis, essential for collagen fibril formation\u003csup\u003e33\u003c/sup\u003e. Despite this, reproducing native oxygen gradients in engineered tissues remains a major challenge.\u003c/p\u003e\n\u003cp\u003eFurthermore, cellular oxygen demand varies significantly across cell types\u003csup\u003e34\u003c/sup\u003e. For instance, hepatocytes consume oxygen at rates of approximately 0.3 nmol/s/10⁶ cells, nearly five times higher than fibroblasts (0.05 nmol/s/10⁶ cells), due to their elevated metabolic activity\u003csup\u003e35\u003c/sup\u003e. This demand also fluctuates depending on cell cycle stage (e.g., stromal vs. terminally differentiated) and cell density within the scaffold. These complex dynamics, combined with the effects of both hypoxia and hyperoxia, underline the necessity for precise oxygen regulation. Notably, short-term hyperoxic pulses can induce protective effects\u0026mdash;a phenomenon known as ischemic preconditioning\u0026mdash;although prolonged exposure can be cytotoxic, underscoring the need for a narrow regulatory window.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOur Approach: Scaling Physicochemical Control to 3D Bioprinted Engineered tissues\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo address these challenges, we implemented controlled cultivation protocols for 3D bioprinted engineered tissues with volumes up to 10.8 cm\u0026sup3; (2\u0026times;2\u0026times;2.7 cm; 12x12 pores). We applied our strategy to a conjunctive tissue composed of human fibroblast\u003csup\u003e36\u003c/sup\u003e and developed generic, robust regulation protocols. A special set-up was here necessary to enable the constant perfusion of the bioprinted engineered tissue with a regulated and monitored cultured media. The experimental system, depicted in \u003cstrong\u003eFigure 2-A\u003c/strong\u003e, feature a commercial regulation vessel equipped with temperature, pH, O\u003csub\u003e2\u003c/sub\u003e and Raman probes, tow gas inlets and a perfusion loop. The 10.8 cm\u0026sup3; tissue was placed in a specially design 3D printed perfusion vessel. Cultures were maintained for 6 days. \u003c/p\u003e\n\u003cp\u003eInitial environmental control strategies were adapted from pharmaceutical bioprocessing protocols useful for cell suspension or microcarrier-based cultures (\u003cstrong\u003eFig. 2-B\u003c/strong\u003e)\u003csup\u003e37\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eTemperature and pH were effectively regulated using this initial protocol, with deviations from the set point of only 0.94% and 0.16%, respectively. In contrast, oxygen regulation proved more challenging (128% deviation from the set point) and required substantial optimisation to produce stable dissolved oxygen level in the perfused culture media. This steady and controlled level is mandatory to overcome the reduced growth rates of fibroblasts (doubling time of 3.5 days in 3D culture, compared to 1.7 days in 2D monolayers) we already demonstrate in bioprinted engineered tissues\u003csup\u003e36\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eInitial regulation relied solely on positive oxygen control (air/O₂ injection in the gas inlet sparger), which proved insufficient since oxygen levels never reached the expected value (set point 35%). To improve regulation, a negative feedback loop using nitrogen (N₂) injection was introduced in the gas inlet sparger, enabling effective oxygen depletion. We further refined control using cascaded PID loop with optimised parameters. This involved a doubling of the proportional term (kP) and a 5-fold decrease of the Integral term (kD), reducing the dissolved oxygen level deviation from 128% to 22% (\u003cstrong\u003eFig.2-C \u003c/strong\u003eand\u003cstrong\u003e \u003cem\u003eSupplementary Table 1\u003c/em\u003e\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eThese changes in the regulation cascade and PID values had an undesirable effect on the pH regulation. This effect was expected since the negative control of the pH is bound to the dissolved CO\u003csub\u003e2\u003c/sub\u003e concentration, which is injected, together with Air and N\u003csub\u003e2\u003c/sub\u003e in the gas mix. The negative oxygen control action was then accompanied by a concomitant negative control action on the pH.\u003c/p\u003e\n\u003cp\u003eFinally, when the set points for pH and dissolved oxygen were set to 7.4 and 40% respectively, both controls were fully successful, with a dissolved oxygen level never lower than the 40% set point and a stable pH of 7.4.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInsights and Implications for Bioprocess Control\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOur results also revealed clear monitoring of oxygen consumption. Indeed, bioprinted conjunctive tissue populated with human fibroblasts showed pronounced oscillations in dissolved oxygen (dO₂) (\u003cstrong\u003eFig. 2-D\u003c/strong\u003e). This aligns with industrial cell culture processes, where dO₂ oscillations are used to estimate oxygen uptake rates (OUR) and reflect both cell number and oxygen consumption rates\u003csup\u003e38\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eThis finding suggests that dO₂ fluctuation patterns could serve as a real-time, non-invasive proxy for estimating cell number and metabolic activity within 3D engineered tissues\u0026mdash;offering, to our knowledge, the first online qualitative description of the respiratory behaviour of tissue-like structures.\u003c/p\u003e\n\u003cp\u003eMoreover, our monitoring system proved highly sensitive to external perturbations. As illustrated in \u003cstrong\u003eFigure 2\u003c/strong\u003e (arrows), media additions and sample withdrawals generated identifiable deviations in the oxygen and pH regulation profiles. Such traceable events provide valuable information on the impact of interventions on the tissue\u0026rsquo;s microenvironment and could inform adaptive feedback in automated culture protocols\u0026mdash;mirroring practices in pharmaceutical manufacturing.\u003c/p\u003e\n\u003cp\u003eFinally, this ability to continuously log and interpret real-time data from the tissue environment is a key step toward creating regulatory-compliant production systems for advanced therapy medicinal products (ATMPs). It establishes a foundational layer for quality-by-design (QbD) approaches in tissue engineering\u003csup\u003e39-41\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMetabolic Activity Monitoring of Engineered Tissues\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMetabolic activity is a core indicator of tissue viability and functionality. Among the many metabolic markers available, lactic acid stands out for its dual relevance: it plays a central role in cellular energy regulation and serves as a sensitive marker of oxygen availability within tissues\u003csup\u003e42\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e43\u003c/sup\u003e. Lactic acid secretion is also associated with major tissues disfunction like lactic acidosis in fatty livers \u003csup\u003e44\u003c/sup\u003e. \u003c/p\u003e\n\u003cp\u003eLactate is also involved in the regulation of major physiological systems\u0026mdash;including cardiovascular, respiratory, and digestive\u0026mdash;and holds clinical significance in the diagnosis and prognosis of a range of conditions\u003csup\u003e45\u003c/sup\u003e. Given its biological importance and interpretability, lactic acid was selected as a demonstrator molecule and assess the feasibility of such online, non-destructive monitoring of metabolic activity within engineered tissues.\u003c/p\u003e\n\u003cp\u003eIn the field of cell culture bioprocessing, several analytical tools have been proposed for real-time metabolic monitoring. Among these, Raman spectroscopy shows great promise for non-invasive, continuous measurement of secreted or consumed metabolites\u003csup\u003e23\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e46\u003c/sup\u003e. In our study, we implemented a bIO-LAB 220 Raman probe directly within the perfusion circuit of the bioreactor vessel (\u003cstrong\u003eFig. 2-A\u003c/strong\u003e), enabling the acquisition of spectra from culture supernatant circulating through the bioprinted engineered tissue (\u003cstrong\u003e\u003cem\u003eSupplementary Figure 3\u003c/em\u003e\u003c/strong\u003e). Unlike many conventional approaches, this setup allowed us to monitor metabolite concentrations without withdrawing samples, offering a real-time and label-free method of assessing tissue metabolic function.\u003c/p\u003e\n\u003cp\u003eOur objective was to demonstrate the versatility of this method across different tissue types and cellular origins. To this end, engineered tissues were bioprinted using cell types from two distinct tissues from human origin, dermis fibroblasts and kidney epithelial cells (HEK). Importantly, lactate concentrations in tissue-based cultures are often in the range of 0.5-2 g.L\u003csup\u003e-1\u003c/sup\u003e \u003csup\u003e36\u003c/sup\u003e, compared to the 3\u0026ndash;12 g.L\u003csup\u003e-1\u003c/sup\u003e commonly encountered in suspension cultures\u003csup\u003e47\u003c/sup\u003e. This is a significant challenge for chemometrics model development, as standard Raman-based mathematical models are generally optimised for higher metabolite concentrations.\u003c/p\u003e\n\u003cp\u003eWe started our study by Raman calibration step. It consists in building a mathematical model correlating enzymatically quantified lactic acid concentrations with Raman spectral data. In order to access to Raman lactic acid signature within complex sample spectra, we had to use chemometric analysis. It involves building Orthogonal Partial Least Squares (OPLS) regression after spectra deconvolution in several principal components. This was performed thanks to Simca\u0026reg; 18 software (Sartorius, Germany). \u003c/p\u003e\n\u003cp\u003eTo build this calibration, we collected 76 discrete culture samples among 7 distinct tissue culture batches along culture which could last up to 32 days. To use best practice for Raman calibration\u003csup\u003e48\u003c/sup\u003e, we also apply sample spiking with added known amount of lactic acid in order to decorrelate its concentrations from other molecules present in the culture samples (ex. Glucose). Then, several spectral pre-processing methods were evaluated and screened \u0026mdash;including baseline correction, normalisation, and first-order derivatives\u0026mdash;to optimise the PLS model fit (R\u0026sup2;\u003csub\u003eCAL\u003c/sub\u003e), its predictive power (Q\u0026sup2;\u003csub\u003eCAL\u003c/sub\u003e), and to minimise root-mean-square errors (RMSE\u003csub\u003eCAL\u003c/sub\u003e). Calibration step enabled to select the four best OPLS models (\u003cstrong\u003eFig. 3-B\u003c/strong\u003e), from which Model 1 calibration curve is presented in \u003cstrong\u003eFigure 3-A\u003c/strong\u003e. \u003c/p\u003e\n\u003cp\u003eTo challenge the predictive performance of these four OPLS calibration models, an external dataset consisting of 21 additional Raman spectra, corresponding to quantified samples from four novel culture tissue batches were used to predict lactic acid. Prediction performance was evaluated by comparison between the real lactic acid quantified concentrations and the predicted concentration. The best predictive PLS model was the one combining spectral treatment with Asymmetric Least Squares (AsLS) baseline correction, Standard Normal Variate (SNV) normalisation, and 1\u003csup\u003est\u003c/sup\u003e derivative transformation. It enabled quantification of lactate concentrations within a range of 0 to 3.5 g.L\u003csup\u003e-1\u003c/sup\u003e, with a prediction precision error RMSE\u003csub\u003ePRED\u003c/sub\u003e of 0.103 g.L\u003csup\u003e-1\u003c/sup\u003e (\u003cstrong\u003eFig. 3-B\u003c/strong\u003e) and a R\u003csup\u003e2\u003c/sup\u003e\u003csub\u003ePRED\u003c/sub\u003e of 0.942. \u003c/p\u003e\n\u003cp\u003eLactic acid Raman quantification was then deployed for the continuous online monitoring of 3D bioprinted conjunctive tissue composed of dermal fibroblasts perfused over a 13-day of culture. The Raman spectra of the perfused medium were acquired every 17 minutes from day 2 corresponding to a total of 933 spectra. The calculated lactate concentration evolution over time is presented in \u003cstrong\u003eFigure 3-C\u003c/strong\u003e. To facilitate lactic acid production trend readability, concentrations were averaged into hourly mean values and represented as black dots (\u003cstrong\u003eFig. 3-C\u003c/strong\u003e). The predicted lactic acid production trend obtained is noisy as the production levels for dermis fibroblasts is very low, i.e. 0.2-0.3 g.L\u003csup\u003e-1\u003c/sup\u003e, compared to the determined prediction performance of 0.1 g.L\u003csup\u003e-1\u003c/sup\u003e. Still, lactic acid secretion trend is consistent with dynamics previously described in static culture for dermal fibroblasts\u003csup\u003e36\u003c/sup\u003e. Additionally at day 6, Raman probe was able to identify the lactic acid dilution implied by 50 mL medium additions, corresponding to 1/3 dilutions (see black arrow on \u003cstrong\u003eFig. 3-C\u003c/strong\u003e). \u003c/p\u003e\n\u003cp\u003eThis proof-of-concept illustrates that Raman spectroscopy can be successfully adapted for online metabolic monitoring in tissue engineering, despite the relatively low metabolite concentrations involved (0.2-0.3 g.L\u003csup\u003e-1\u003c/sup\u003e).These findings echo its established use in industrial-scale bioprocessing for molecules such as glucose, amino acids, and secreted proteins\u003csup\u003e23\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e45\u003c/sup\u003e. Although the development of a robust calibration model still requires a representative dataset of at least 75-100 samples\u0026mdash;sufficient in both size and variability\u0026mdash;the approach shows high potential for broader applications in engineered tissue monitoring. Going forward, the ability to monitor nutrient consumption and by-product accumulation in real time offers not only an advanced level of bioprocess control, but also a new window into tissue physiology and metabolic dynamics. Expanding this platform to include additional analytes and tissue types could pave the way for the creation of intelligent bioreactor systems capable of adapting in real time to the evolving needs of the engineered tissue.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMorphology Monitoring of Engineered Tissues\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAs engineered tissues mature \u003cem\u003ein vitro\u003c/em\u003e, their internal structures undergo continuous remodelling driven by cellular activity. These changes include extracellular matrix (ECM) secretion and remodelling, formation of vessels and cavities, compaction, and overall densification of the tissue\u003csup\u003e49\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e50\u003c/sup\u003e. As a result, the geometry of 3D bioprinted engineered tissues is expected to evolve significantly throughout the culture period and must be monitored over time to ensure fidelity to the original design and to understand tissue development dynamics.\u003c/p\u003e\n\u003cp\u003eTo date, most techniques available for characterizing internal morphology are either destructive (histology or electron microscopy), require extensive sample preparation (light-sheet or confocal microscopy)\u003csup\u003e51\u003c/sup\u003e or lack depth of analysis (Optical coherence tomography)\u003csup\u003e28\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e52\u003c/sup\u003e. These constraints limit the ability to perform real-time, longitudinal monitoring of tissue morphology throughout maturation.\u003c/p\u003e\n\u003cp\u003eMagnetic Resonance Imaging (MRI), widely recognized \u003cem\u003ein vivo\u003c/em\u003e for its safety and non-invasiveness, offers a promising solution for the non-destructive monitoring of 3D engineered tissues\u003csup\u003e53\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e54\u003c/sup\u003e. In this study, we evaluated the application of MRI for tracking the morphological evolution of a 3D bioprinted engineered tissue seeded with human fibroblasts over a 15-day culture period.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFigure 4\u003c/strong\u003e presents MRI data acquired using a 7 Tesla system with T2-weighted sequences, chosen for their high sensitivity to water content and thus their ability to distinguish hydrogels from surrounding fluids and matrix. High-resolution imaging was achieved without the use of any contrast agent, yielding isotropic voxels of 137\u0026times;137\u0026times;137 \u0026micro;m\u0026sup3;. As shown in \u003cstrong\u003eFigure 4-A\u003c/strong\u003e, MR images acquired immediately after bioprinting offered clear visualization of the entire 10.8 cm\u0026sup3; engineered tissue, with excellent contrast between voids and matrix. Multiplanar views (sagittal, coronal, and axial) allowed comprehensive assessment of internal geometries, including nascent cavities.\u003c/p\u003e\n\u003cp\u003eTo quantitatively assess morphological fidelity, the segmented MRI data were compared with the original bioprinting STL design file (\u003cstrong\u003eFig. 4-B\u003c/strong\u003e). This analysis, conducted using 3DSlicer (USA), compared over one million discrete spatial points across both models. Results showed that 87.6% of the MRI-derived tissue geometry deviated by less than 1 mm (representing ~4% of the engineered tissue\u0026apos;s average dimension) from the original STL. These minor discrepancies are primarily attributed to extrusion-based bioprinting resolution limits and handling-induced deformations, rather than limitations of the imaging process itself\u003csup\u003e55\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eTo demonstrate longitudinal monitoring capabilities, the same engineered tissue was imaged after 16 days of culture under controlled physicochemical conditions (\u003cstrong\u003eFig. 4-C\u003c/strong\u003e). MRI revealed significant morphological evolution, including pore occlusion (green star), shape deformation, and localized matrix degradation (red arrows). These features were corroborated by histological analysis of corresponding tissue sections (Masson\u0026rsquo;s trichrome staining, \u003cstrong\u003eFig. 4-D\u003c/strong\u003e), providing strong correlation between non-destructive and classical destructive approaches. Histology results also permit the identification of deposited extracellular matrix (green coloration) and embedded cells (black arrows), after 16 days of culture.\u003c/p\u003e\n\u003cp\u003eTaken together, these results validate MRI as a reliable, non-destructive modality for high-resolution morphological monitoring of 3D bioprinted engineered tissues. The ability to acquire quantitative and longitudinal data without the need for labelling or sample preparation represents a major advancement in tissue engineering. This approach enables dynamic tracking of tissue maturation and deformation, contributing valuable insights for engineered tissue design optimization, quality control and regulatory documentation, all essential steps for advancing regenerative medicine applications.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePerfusion Monitoring of Engineered Tissues\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAs engineered tissues increase in volume and complexity, particularly beyond the cubic centimetre scale, simple diffusion mechanisms become insufficient to sustain cellular viability across the entire engineered tissue\u003csup\u003e31\u003c/sup\u003e. In such cases, active perfusion of culture media is required to ensure the homogeneous delivery of oxygen and nutrients while facilitating the removal of metabolic waste. This challenge is particularly relevant in the context of engineered tissues exceeding 10 cm\u0026sup3;, such as those investigated in the present study. To design effective perfusion strategies and verify their implementation, we established a dual approach combining numerical simulation and experimental mapping of internal flow fields: computational fluid dynamics (CFD) was used to predict perfusion performance, and MRI velocimetry\u003csup\u003e56\u003c/sup\u003e was applied to non-destructively quantify fluid flow through the engineered tissues.\u003c/p\u003e\n\u003cp\u003eWe first employed CFD simulations using ANSYS Fluent to evaluate nutrient flow within the tissue perfusion vessel (\u003cstrong\u003eFig. 2-A\u003c/strong\u003e) housing a 12\u0026times;12 channels mock scaffold, 3D printed in PLA. Simulations were performed at two different flow rate, 2 mL.min\u003csup\u003e-1\u003c/sup\u003e and 20 mL.min\u003csup\u003e-1\u003c/sup\u003e. The resulting velocity maps, in the vertical plan including the inlet and outlet, are presented in \u003cstrong\u003eFigure 4-A\u003c/strong\u003e. A clearly heterogeneous distribution of the perfused liquid velocity was evidenced for both flow rates, consequence of the preferential horizontal flow path within the scaffold which presented lateral pores of 1.2x1.2 mm and vertical pores of 1.2x0.4 mm (\u003cstrong\u003e\u003cem\u003eSupplementary Figure 1\u003c/em\u003e\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eOnce these simulation performed, flow-sensitive MRI was used to experimentally measure internal fluid dynamics within the 12\u0026times;12 channels mock scaffold subjected to identical flow conditions, i.e. 2 mL/min and 20 mL/min. In the particular case of this flow-sensitive MRI experiment, the previously used 7 Tesla MRI scanner was set up to achieve an isotropic resolution of 1 mm\u0026sup3;. The system was tuned to detect flow velocities ranging from 200 \u0026micro;m.sec\u003csup\u003e-1\u003c/sup\u003e to 10 cm.sec\u003csup\u003e-1\u003c/sup\u003e, and imaging was performed without the use of any contrast agent. \u003cstrong\u003eFigure 4-C\u003c/strong\u003e present the velocity maps, in the vertical plan including the inlet and outlet, for the two tested flow rates. Strong agreement was observed in terms of global flow direction, velocity magnitude and pattern, when comparing MRI results to the CFD simulations. However, several discrepancies emerged mainly due to the presence of air bubbles in the experimental set-up. These bubbles, easily identified in the morphological image (white arrows in \u003cstrong\u003eFig. 4-B\u003c/strong\u003e) had a strong impact on the flow distribution.\u003c/p\u003e\n\u003cp\u003eMRI-based velocimetry also enabled the visualisation of perfusion flows in 3D. Examples are given in \u003cstrong\u003eFigure 4-D\u003c/strong\u003e with 2 different views of the flow path volumes at the 2 different flow rates. Even though these representation are difficult to appreciate in 2D, their observation as 3D animation are of great interest to appreciate the internal flow geometries (\u003cstrong\u003e\u003cem\u003eSupplementary Video 1 \u003c/em\u003e\u003c/strong\u003eand\u003cstrong\u003e\u003cem\u003e 2\u003c/em\u003e\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eFinally, the complementary use of CFD and MRI provides a powerful framework for understanding and optimising perfusion in bioengineered tissues. CFD offers a rapid, flexible means of screening engineered tissue designs and bioreactor parameters \u003cem\u003ein silico\u003c/em\u003e, whereas MRI yields spatially resolved, empirical data that can be acquired non-destructively and applied throughout the culture process. Importantly, MRI-based velocimetry is directly transferable to living engineered tissues and could be used to track changes in perfusion as a function of tissue maturation, cavity formation, or vascularisation.\u003c/p\u003e\n\u003cp\u003eThis combined approach lays the foundation for future development of feedback-regulated perfusion systems, capable of dynamically adjusting flow rates in response to tissue metabolic needs. Such systems would represent a significant advancement in the field of bioprocess control, enabling safer and more efficient maturation of large-scale engineered tissues for clinical and pharmaceutical applications.\u003c/p\u003e"},{"header":"Discussion and Conclusion","content":"\u003cp\u003eThe increasing sophistication of biofabricated tissues, both in scale and biological complexity, demands a corresponding evolution in analytical strategies. As demonstrated in this study, non-destructive, real-time monitoring of engineered tissues is not only technically feasible but also essential for controlling the dynamic and multiscale processes underlying tissue maturation. We showed that the physicochemical environment, structural integrity, internal perfusion, and metabolic activity of 3D bioprinted engineered tissues can all be monitored using integrated sensor-based and imaging tools commonly available in pharmaceutical bioprocessing and clinical imaging, yet underexplored in tissue engineering contexts.\u003c/p\u003e\u003cp\u003eThe application of \u003cb\u003eRaman spectroscopy\u003c/b\u003e for monitoring lactic acid secretion from engineered tissues constitutes a significant advance in metabolic surveillance. Through PLS modelling and spectrum pre-processing, we quantified lactate at physiologically relevant concentrations (from 0 to 3.5 g.L\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e) with a resolution of 0.10 g.L\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, enabling temporal mapping of metabolic activity every 17 minutes. This continuous readout correlates with known proliferation phases and provides a framework for applying metabolic control in future smart bioreactor platforms.\u003c/p\u003e\u003cp\u003eOur work also highlights the critical importance of \u003cb\u003eoxygen regulation\u003c/b\u003e in volumetric cultures. Standard cell culture protocols proved inadequate for large-scale engineered tissues, leading to significant deviations from target dissolved oxygen concentrations and reduced cell growth. By implementing dual-gas control and optimising PID parameters, we achieved a substantial improvement in oxygen regulation, reducing deviations from 131\u0026ndash;11.5% in fibroblast culture. These improvements translated into better environmental stability and, likely, more reproducible biological outcomes.\u003c/p\u003e\u003cp\u003eIn terms of \u003cb\u003emorphological monitoring\u003c/b\u003e, 7 Tesla MRI allowed for contrast agent-free, longitudinal tracking of 3D engineered tissues with a resolution sufficient to capture structural changes over time. MRI-derived geometries closely matched the original STL files (with over 87% of points deviating by less than 1 mm), and later scans revealed remodelling events such as pore closure and matrix compaction\u0026mdash;validated by histology. This non-destructive approach offers valuable insights into tissue integrity and deformation throughout culture.\u003c/p\u003e\u003cp\u003eFinally, our dual strategy for assessing \u003cb\u003etissue perfusion\u003c/b\u003e\u0026mdash;via CFD simulation and MRI velocimetry\u0026mdash;proved essential for validating flow distribution. Numerical models provided fast, scalable design tools, while MRI confirmed internal velocity profiles in printed engineered tissues, capturing flow heterogeneity as a function of architecture and flow rate. This capability supports the implementation of adaptive perfusion systems that can evolve with the tissue\u0026rsquo;s metabolic needs.\u003c/p\u003e\u003cp\u003eTogether, these tools outline a modular, scalable, and highly transferable analytical framework for engineered tissue monitoring. By adapting technologies already validated in pharmaceutical manufacturing, this study bridges a critical gap between traditional tissue culture and the emerging field of automated, closed-loop biofabrication. Beyond improving process control, these monitoring strategies lay the groundwork for regulatory alignment, standardisation, and quality assurance, both key steps toward the clinical translation of engineered tissues.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cem\u003eCell culture \u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eTwo cell types were chosen for this study: human dermal fibroblasts (obtained from Hospice Civil Lyon cell bank, France) and human kidney cells (HEK293T, provided by Dr. C. Maisse-Paradisi, INRA-UCBL-EPHE \u0026quot;Viral Infections and Comparative Pathology\u0026quot;, Lyon, France).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFibroblasts and HEK293T were cultured in Dulbecco\u0026apos;s Modified Eagle Medium (DMEM) (Thermo Fisher Scientific, 31966-021) supplemented with 10\u0026nbsp;% (v/v) FBS (Gibco Cell Culture, 10270-106). Cells were pre-cultured at 5\u0026nbsp;% CO\u003csub\u003e2\u003c/sub\u003e, 80 % humidity and 37\u0026deg;C before bioprinting. Fibroblasts and HEK293T were passaged once and twice a week, respectively.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e3D Bioprinting of engineered tissues\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e3D bioprinting bioink was formulated using bovine gelatine (G1890 Sigma, France), very low viscosity alginate (A18565.36 Alfa Aesar, Thermo Fisher France) and fibrinogen from bovine plasma (F8630 Sigma, France). All components were handle under sterile laminar flow to ensure sterility. Stock solution of 0.2 g/mL gelatine, 0.04 g/mL alginate and 0.08 g/mL fibrinogen were dissolved, without any stirring for 18 hours at 37 \u0026deg;C, in DMEM (without calcium, with glutamax-1, Invitrogen, France) supplemented with 10% foetal calf serum (HyClone, USA), 20 \u0026micro;g/ml gentamicin (Pantapharm, France), 100 UI/ml penicillin/streptomycin (Sarbach, France) and 1 \u0026micro;g/ml amphotericin B (Bristol Myers Squibb, France).\u003c/p\u003e\n\u003cp\u003eFor bioink preparation, trypsinated cells were first suspended in calcium-free DMEM supplemented with 10 % FBS and enumerated. Targeted cell concentration was adjusted by pelleting the appropriate number of cells at 300 g (Fibroblasts, HEK293T) for 5\u0026nbsp;min. The pellet was suspended in the proper volume of\u0026nbsp;0.08 g/mL fibrinogen solution to reach 2.2 x 10\u003csup\u003e5\u003c/sup\u003e cells/mL for fibroblast and 2.6 x 10\u003csup\u003e6\u003c/sup\u003e cells/mL for HEK 293T.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThen, to this cell suspension in fibrinogen, the appropriate volumes of alginate and gelatine stock solutions were added in order to reach a final composition of 0.02 mg/mL fibrinogen, 0.02 mg/mL\u003csub\u003e\u0026nbsp;\u003c/sub\u003ealginate and 0.05 mg/mL\u003csub\u003e\u0026nbsp;\u003c/sub\u003egelatine.\u0026nbsp;The bioink was homogenized and incubated for 15 min at 37\u0026deg;C.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAfter homogenization, sterile cartridge (Nordson EFD, France) was filled with the bioink and incubated 30 minutes at 21\u0026deg;C to stabilise the bioink rheological properties. The cartridge was then loaded in a 6-axis robotic bioprinter (BioAssemblyBot\u0026reg;, Advanced Lifescience Solutions, USA) and used to print\u0026nbsp;10.8 cm\u003csup\u003e3\u003c/sup\u003e parallelepiped macroporous 3D structures (2x2x2.7 cm, 12x12 pores, internal pore dimensions of 1.2x1.2 mm).\u003c/p\u003e\n\u003cp\u003eA 800 \u0026micro;m diameter, 6.35 mm long needle (Nordson EFD, France) was used to bioprint at a set speed an pressure of 8 mm/sec and 20-50 Psi, respectively.\u003c/p\u003e\n\u003cp\u003eOnce bioprinted, the tissues were consolidated with a solution composed of 40 mg/mL transglutaminase (TAG) (ACTIVA WM - Ajinomoto), 10 U/mL thrombin from bovine plasma (T4648-10KU - Merck) and 270 mM CaCl\u003csub\u003e2\u003c/sub\u003e (C5670-500G - Merck). The consolidation process was carried out at 37\u0026deg;C for 2 hours. The cellularised structures were then rinsed twice with sterile physiological serum (Versol, France) before being introduced in the tissue perfusion vessel.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eEngineered tissues perfusion chamber design and production\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eComputer Assisted Design (CAD) of the tissue perfusion chamber was performed using Autodesk\u0026reg; Fusion 360\u0026trade;. CAD files, available in a GitHub repository (freely available at https://github.com/FabricAdvancedBiology/Multiflow_cell) were converted to .STL format to be transferred to an Object30 Pro inkjet printer (Stratasys, USA). The chamber was 3D-printed using VeroClear resin (Stratasys, USA). Once printed, the support material was removed using a high pressure waterjet (Stratasys). The 3D-printed chamber was then incubated overnight in 70 % ethanol at room temperature to remove all leachable compounds, and then finally rinsed in milliQ water. Prior any use for tissue perfusion, the chamber was steam sterilized at 120\u0026deg;C, 2 Bars for 20 minutes.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eSet up for physicochemical environment monitoring\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eA controlled and regulated physicochemical tissue environment was established using a commercial bioreactor (Eppendorf DASbox\u0026reg; 250mL) connected to the 3D-printed tissue perfusion chamber through a peristaltic pumping systems.\u003c/p\u003e\n\u003cp\u003eTo monitor and control the\u0026nbsp;physicochemical environment\u003cem\u003e\u0026nbsp;\u003c/em\u003eparameters (pH, dissolved oxygen, temperature), a custom Cytosys controller (Ipratech, Belgium) was employed. Customization of the system included an Applikon heating jacket 30W MINIBIO 250 heating blanket (Applikon, Sweden), a custom stirrer motor, 12 mm pH and oxygen probes (Hamilton) and two temperature probes (Pt100). The initial dissolved oxygen level regulation loop involved air and nitrogen injections, while pH was regulated thanks to CO\u003csub\u003e2\u003c/sub\u003e and 0.2 M NaOH solution.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eBefore cultivation, Eppendorf DASbox\u0026reg; bioreactors was steam sterilised. A day before its connexion to the tissue perfusion chamber, the bioreactor was filled with culture medium and the pH, temperature and oxygen monitoring/regulation started to stabilise the probes\u0026rsquo; response before culture.\u003c/p\u003e\n\u003cp\u003eOnce bioprinted and consolidated, the 10.8 cm\u003csup\u003e3\u003c/sup\u003e engineered tissues\u003cem\u003e\u0026nbsp;\u003c/em\u003ewere sterilely inserted, under laminar flow, into the 3D-printed perfusion chamber. The tissue perfusion chamber was sealed, connected to perfusion tubing and the perfusion peristaltic pump started to generate perfusion at 2 or 20 mL/min perfusion rates.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eLactic acid enzymatic assay\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eSpent media were recovered daily through sampling, aliquoted in triplicate and stored at -20\u0026deg;C. These samples were further used to quantify the lactic acid produced by cell metabolism using the L-Lactic Acid Assay from Megazyme (L-Lactic Acid Assay Kit, K-LATE, Megazyme). The assay was performed in 96-well plates with an autosampler procedure according to the manufacturer\u0026apos;s instructions. Each sample was analysed in triplicate at 340 nm with a spectrophotometer (TECAN infinite\u003csup\u003e\u0026reg;\u003c/sup\u003e). Samples presenting lactate concentration above the quantification linearity were diluted in deionised water. All results were normalised with a fresh culture medium used as a blank.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eOnline Raman spectroscopy\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe acquisition of culture medium Raman spectra was performed \u003cem\u003eonline\u003c/em\u003e with an Rxn2 IOT Raman spectroscopy analyser (Endress Hauser) and associated Raman probe (bIO-LAB 220; Endress Hauser) kindly provided by Kaiser-Endress. The probe was placed within the DASbox\u0026reg; tank before set-up sterilisation. The excitation wavelength of the laser was 785 nm. A preliminary blank spectrum was acquired and lasted 15 minutes. The acquisition parameters were optimised to target between 50 to 80% of signal saturation. Thus, acquisition were set at 40 s with 20 accumulation counts thus corresponding to 17 minutes accumulation in total. After Raman spectrum acquisition, \u003cem\u003eonline\u003c/em\u003e lactate monitoring was calibrated based on \u003cem\u003eoffline\u003c/em\u003e samples. 76 off-line spectra were generated from spiked fresh or spent medium samples from fibroblast\u0026apos;s culture with concentrated Lactic acid solution.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eMorphology monitoring of engineered tissues by Magnetic Resonance Imaging (MRI)\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eAt the end of the culture, engineered tissues were fixed by incubation in AntigenFix (DiaPath, P0014) overnight at 4\u0026deg;C. The engineered tissues were then rinsed and kept in ethanol 70 % before MRI analysis. The MRI morphology protocol was carried out on a 7 Tesla Bruker BioSpec MR system (Bruker Biospin GbmH, Germany) equipped with a 400 mT/m maximal amplitude gradient set and controlled using a Bruker workstation interfaced with ParaVision5.1 software for data acquisition and post-processing. A transmit-receive radio-frequency body coil (outer diameter 112 mm and inner diameter 72 mm) was employed for \u003cem\u003ein vitro\u003c/em\u003e MR image acquisition. To acquire 3D high-resolution MR images of the engineered tissues, a tridimensional T2-weighted MR sequence based on the Rapid Acquisition with Relaxation Enhanced (RARE) method was performed on transversal orientation. The acquisition parameters were as follows; Echo Time (TE) 35.8 ms, Repetition Time (TR) 600 ms, bandwidth 75 kHz, RARE accelerator factor 8, and average number 2. A total of 256 slices of 273 \u0026micro;m thickness were acquired within a field of view of 3.5 x 3.5 x 3.5 cm\u003csup\u003e3\u003c/sup\u003e, and an in-plane\u0026nbsp;matrix size of 256x256x128\u0026nbsp;interpolated to 256x256x256 pixels, providing a final in-plane isotropic resolution\u0026nbsp;of\u0026nbsp;137x137x137 \u0026micro;m\u003csup\u003e3\u003c/sup\u003e. The total acquisition time was 1h1m26s.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003ePerfusion Monitoring of Engineered Tissues\u0026nbsp;\u003c/em\u003e\u003cem\u003eby Magnetic Resonance Imaging\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eVelocimetry Magnetic Resonance Imaging (MRI) protocol was carried out on a 7 Tesla Bruker BioSpec MR system. Mock scaffolds were 3D printed thanks to a Prusa MINI+ printer (PRUSA Research, Czech Republic) using 0.1 mm resolution slicing and a PLA filament (Prusament, PRUSA Research, Czech Republic). The structure of the mock scaffold is presented in \u003cstrong\u003e\u003cem\u003eSupplementary Figure 1\u003c/em\u003e\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eThey were sequentially inserted in the bioreactor\u0026rsquo;s chamber, leaving a gap of 2.2 mm towards the outlet wall and 2.4 mm towards 1 lateral wall. The bioreactor was then sealed and inserted inside the transmit-receive body coil positioned in the centre of the 7T MRI system (\u003cstrong\u003e\u003cem\u003eSupplementary Figure\u003c/em\u003e 2\u003c/strong\u003e). A flow loop was set up using a peristaltic pump, Masterflex Ismatec ISM834C. The whole set-up, including the rollers\u0026rsquo; stretch, was connected with an autoclavable translucent hose, Versilic flexible tube, \u0026Oslash; 3mm x \u0026Oslash; 6mm (228-1194-VWR). The pump speed was calibrated by mapping the display reading to the displaced weighted volume, sweeping the whole range separated by 6 sampled speeds. A linear regression model was established (coefficient of determination R\u003csup\u003e2\u003c/sup\u003e = 0.993). The hydraulic circuit was filled with stained water. The incoming flow hose was connected to one bottom orifice, while the outgoing flow hose was released in the opposite-side upper orifice. The pump was set to run the experiments with two different pump flows: 2 mL/min and 19.6 mL/min.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe velocity map sequence acquisition was carried out using the flow-imaging technique known as Flow Map, which is based on the phase contrast method \u003csup\u003e22\u003c/sup\u003e. Phase contrast MRI (PC-MRI) is a non-invasive imaging technique that does not require contrast agents and can provide high-resolution images of blood flow in any direction. PC MRI relies on the principle that the phase of the MRI signal is proportional to the velocity of the protons. To perform PC-MRI, two images are acquired with oppositely applied magnetic field gradients. This creates a phase difference between the two images that is proportional to the velocity of the protons. The phase difference image can then be used to calculate the velocity of the flowing fluid. The sequence was set with the following parameters: average of four images, eco time: 5 ms, repetition time: 50 ms. The maximum flow velocity measured with a given flow imaging sequence was set at 20 cm.s\u003csup\u003e-1\u003c/sup\u003e. In comparison, the minimum flow velocity measured with the given pulse sequence parameters was set at 0.74 cm.s\u003csup\u003e-1\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eFour scenarios, two velocities for two macroporous structures, were acquired sweeping consecutive sagittal planes (XY) separated by 1mm and encoding the velocity vector components in all directions (XYZ). Each acquisition was set to cover a field of view of 5 x 5 cm, represented in 256 x 256 pixel images with a resolution of 195 \u0026mu;m.pixel\u003csup\u003e-1\u003c/sup\u003e for each direction (XYZ). The photos were encapsulated both in DICOM and 2D sequence formats. Each experiment was completed in 58 min. Morphological images were also acquired in the same sweeping positions applying the T2-weighted spin-echo sequence, with and without running flow.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003ePost-processing treatment of velocimetry data\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe raw data captured in the binary file format owned by BRUKER, namely, 2D sequence data (.2dseqs) and then it was decoded, using the BRUKER\u0026rsquo;s Application Programming Interface (API) installed in the programming language Python and converted from a four-dimensional array to segregated images in text file format. The 92 images corresponding to the 23 cuts were loaded in a MATLAB to be post-processed. The Euclidean norm of the vector velocity was calculated by combining the vector components in the three directions and scaling the result to obtain a value in meters per second. All the images were inspected to spot the bubbles\u0026apos; cross-sections, characterized by a region of high heterogeneity of color (spatial noise). Each image was treated individually to take the values in the bubble region to 0 m.sec\u003csup\u003e-1\u003c/sup\u003e. The morphological images were used to create a binary mask to remove the noise from the periphery of the liquid chamber.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe 3-dimensional representation of the velocity norm in all the volume was obtained by upsampling the images in the Z direction using the interpolation protocol defined in [10.1038/srep29936] to even out the resolution to 195 \u0026mu;m/slide and keep the real proportions. The new set of 110 images was stacked and rendered in the software 3D slicer using a smoked appearance to maximize the overall view in one shot.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eHistological analyses\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eFixed engineered tissues, stored in 70% ethanol were paraffin embedded using successive incubations at 65\u0026deg;C in 95% ethanol, 100% ethanol, ethanol/butanol mix, and butanol/methylcyclohexane mixes. Each incubation lasted for 20 min. Paraffin impregnation was then performed at 56\u0026deg;C for 2 h. \u0026nbsp;Paraffin blocks were formed using metal moulds. The dehydrated engineered tissues were sliced into thin sections of 5 \u0026micro;m using a microtome. The sections were disposed of on Superfrost slides (Epredia, 12321). The sections were stained for Masson\u0026rsquo;s Trichrome to observe cell and collagen repartition inside the structure. Sections were stained with Groat\u0026rsquo;s hematoxylin for 1 min, Fuschin-ponceau for 3 min, G orange-phosphomolybdic acid for 3 min, and fast green for 3 min. Each staining step was followed by a water or 1% acetic acid wash. Sections were then dehydrated by rapid immersions in 70%, 95%, and 100% ethanol, and finally, slides were mounted with ExPert Mounting Medium (CellPath, SEA-1604-00A).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eComputational fluid dynamics (CFD)\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eFluid dynamics simulations were performed using the commercial finite volume-solver ANSYS Fluent2019. Based on the previous work of Pourchet et al. \u003csup\u003e23\u003c/sup\u003e, a single-phase with laminar flow was supposed, and a steady-state approach was applied for the solving of both continuity and momentum Navier-Stokes equations (Eq. (1) and (2)), with \u003cem\u003e⍴\u003c/em\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003ethe fluid density, \u003cstrong\u003ev\u003c/strong\u003e the fluid velocity, \u003cem\u003ep\u003c/em\u003e the pressure, \u003cstrong\u003e𝜏\u003c/strong\u003e the Reynolds stress tensor and\u003cstrong\u003e\u0026nbsp;\u003cem\u003eg\u003c/em\u003e\u003c/strong\u003e the acceleration due to gravity. Fluid properties were like those of water at 37\u0026deg;C, namely\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cem\u003e\u0026rho;\u003c/em\u003e = 993 kg m\u003csup\u003e-3\u003c/sup\u003e and dynamic viscosity \u003cem\u003e\u0026micro;\u003c/em\u003e = 6.92 \u0026times; 10\u003csup\u003e-4\u003c/sup\u003e Pa.s.\u003c/p\u003e\n\u003cp\u003e\u003cimg src=\"data:image/png;base64,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\" width=\"823\" height=\"95\"\u003e\u003c/p\u003e\n\u003cp\u003eParameters discretisation was achieved by a mesh divided into approximately 19 million tetrahedral meshes. The perfusion was performed by setting a fixed flow rate and pressure outlet corresponding to the flow inlet/outlet. All other boundaries were set as no-slip walls. Finally, solving equations used the SIMPLE method for the pressure-velocity coupling and 2\u003csup\u003end-\u003c/sup\u003eorder UPWIND schemes for pressure and momentum transport equations. Simulations were run until both a minimum criterion on residuals (\u0026lt; 10\u003csup\u003e-3\u003c/sup\u003e) and stabilisation of the fluid velocity, the pressure outlet and the wall shear stress were reached. CFD post-treatment was performed using ANSYS CFD-Post software. According to the tissue internal flow monitoring, specific velocity flow fields were plotted to compare the experimental data with the modelled ones.\u0026nbsp;\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eFunding Declaration\u003c/h2\u003e\n\u003cp\u003eThe authors declare no funding relate to the present study.\u003c/p\u003e\n\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\n\u003cp\u003eLC, SP, NEK, EP, SL, and CL realised the experiments and generated the data. CM, YG, SL, KM, CL, and EP performed data curing, analysis, model building, and figure building. EP, CM, YG, and CL participated in the manuscript writing. All authors reviewed the manuscript.\u003c/p\u003e\n\u003ch2\u003eAcknowledgement\u003c/h2\u003e\n\u003cp\u003eWe express our sincere gratitude to Jean Lynce Gnanago for his pivotal role in initiating the research on MRI flow analysis during his PhD. His dedicated study forms the cornerstone of the work presented in this paper. We acknowledge Radu Bolbos and CERMEP MRI 7T for their invaluable support in facilitating the MRI acquisition for this study\u0026mdash;special thanks to the Sartorius PAT Team for their crucial assistance in Raman\u0026apos;s chemometric analysis. We also acknowledge Sartorius for their financial support, which contributed to completing this work, and for providing the SIMCA software. We also acknowledge Kaiser Endress-Hauser and, in particular, Thomas Perilli for the loan of Raman equipment Rxn2 and the associated probes. The collaborative efforts and support from these individuals and organisations have been instrumental in the successful execution and completion of this research project.\u003c/p\u003e\n\u003ch2\u003eData Availability\u003c/h2\u003e\n\u003cp\u003eData will be available upon request to the corresponding author.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003ePourchet, L.\u003cem\u003e et al.\u003c/em\u003e Large 3D bioprinted tissue: Heterogeneous perfusion and vascularization. \u003cem\u003eBioprinting\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, (2019).\u003c/li\u003e\n\u003cli\u003eMarquette, C. A.\u003cem\u003e et al.\u003c/em\u003e Unlocking the potential of bio-inspired bioinks: A collective breakthrough in mammalian tissue bioprinting. \u003cem\u003eBioprinting\u003c/em\u003e \u003cstrong\u003e41\u003c/strong\u003e, (2024).\u003c/li\u003e\n\u003cli\u003eLi, H.\u003cem\u003e et al.\u003c/em\u003e Digital light processing (DLP)-based (bio)printing strategies for tissue modeling and regeneration. \u003cem\u003eAggregate\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, (2023).\u003c/li\u003e\n\u003cli\u003eLi, W.\u003cem\u003e et al.\u003c/em\u003e Stereolithography apparatus and digital light processing-based 3D bioprinting for tissue fabrication. \u003cem\u003eiScience\u003c/em\u003e \u003cstrong\u003e26\u003c/strong\u003e, (2023).\u003c/li\u003e\n\u003cli\u003eShiwarski, D. J.\u003cem\u003e et al.\u003c/em\u003e Emergence of FRESH 3D printing as a platform for advanced tissue biofabrication. \u003cem\u003eAPL Bioeng\u003c/em\u003e \u003cstrong\u003e5\u003c/strong\u003e, (2021).\u003c/li\u003e\n\u003cli\u003eBernal, P. N.\u003cem\u003e et al.\u003c/em\u003e Volumetric Bioprinting of Complex Living-Tissue Constructs within Seconds. \u003cem\u003eAdv Mater\u003c/em\u003e \u003cstrong\u003e31\u003c/strong\u003e, (2019).\u003c/li\u003e\n\u003cli\u003eZhang, Y. S.\u003cem\u003e et al.\u003c/em\u003e 3D extrusion bioprinting. \u003cem\u003eNature Reviews Methods Primers\u003c/em\u003e \u003cstrong\u003e1\u003c/strong\u003e, (2021).\u003c/li\u003e\n\u003cli\u003eDuarte, C. D. F.\u003cem\u003e et al.\u003c/em\u003e Synchronized Dual Bioprinting of Bioinks and Biomaterial Inks as a Translational Strategy for Cartilage Tissue Engineering. \u003cem\u003e3d Print Addit Manuf\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e, (2019).\u003c/li\u003e\n\u003cli\u003eAli, A. S. M.\u003cem\u003e et al.\u003c/em\u003e 3D bioprinting of liver models: A systematic scoping review of methods, bioinks, and reporting quality. \u003cem\u003eMaterials Today Bio\u003c/em\u003e \u003cstrong\u003e26\u003c/strong\u003e, (2024).\u003c/li\u003e\n\u003cli\u003eFang, Y.\u003cem\u003e et al.\u003c/em\u003e 3D printing of vascularized hepatic tissues with a high cell density and heterogeneous microenvironment. \u003cem\u003eBiofabrication\u003c/em\u003e, (2023).\u003c/li\u003e\n\u003cli\u003eGrigoryan, B.\u003cem\u003e et al.\u003c/em\u003e Multivascular networks and functional intravascular topologies within biocompatible hydrogels. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e364\u003c/strong\u003e, (2019).\u003c/li\u003e\n\u003cli\u003eLee, A.\u003cem\u003e et al.\u003c/em\u003e 3D bioprinting of collagen to rebuild components of the human heart. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e365\u003c/strong\u003e, (2019).\u003c/li\u003e\n\u003cli\u003eSkylar-Scott, M. A.\u003cem\u003e et al.\u003c/em\u003e Biomanufacturing of organ-specific tissues with high cellular density and embedded vascular channels. \u003cem\u003eSci Adv\u003c/em\u003e \u003cstrong\u003e5\u003c/strong\u003e, (2019).\u003c/li\u003e\n\u003cli\u003eLee, S. J.\u003cem\u003e et al.\u003c/em\u003e 3D Bioprinting for Engineered Tissue Constructs and Patient-Specific Models: Current Progress and Prospects in Clinical Applications. \u003cem\u003eAdv Mater\u003c/em\u003e \u003cstrong\u003e36\u003c/strong\u003e, (2024).\u003c/li\u003e\n\u003cli\u003eEdri, S.\u003cem\u003e et al.\u003c/em\u003e 3D Bioprinting of Multicellular Stem Cell-Derived Constructs to Model Pancreatic Cell Differentiation. \u003cem\u003eAdv Func Mater\u003c/em\u003e \u003cstrong\u003e34\u003c/strong\u003e, (2024).\u003c/li\u003e\n\u003cli\u003eScognamiglio, C.\u003cem\u003e et al.\u003c/em\u003e Bioprinting stem cells: building physiological tissues one cell at a time. \u003cem\u003eAmerican Journal of Physiology-Cell Physiology\u003c/em\u003e \u003cstrong\u003e319\u003c/strong\u003e, (2020).\u003c/li\u003e\n\u003cli\u003eLeferink, A. M.\u003cem\u003e et al.\u003c/em\u003e Methods of Monitoring Cell Fate and Tissue Growth in Three-Dimensional Scaffold-Based Strategies for In Vitro Tissue Engineering. \u003cem\u003eTissue Engineering Part B: Reviews\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, (2016).\u003c/li\u003e\n\u003cli\u003eMather, M. L.\u003cem\u003e et al.\u003c/em\u003e Meeting the Needs of Monitoring in Tissue Engineering. \u003cem\u003eRegenerative Medicine\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, (2007).\u003c/li\u003e\n\u003cli\u003eHolzberg, T. R.\u003cem\u003e et al.\u003c/em\u003e Sensors for biomanufacturing process development: facilitating the shift from batch to continuous manufacturing. \u003cem\u003eCurrent Opinion in Chemical Engineering\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, (2018).\u003c/li\u003e\n\u003cli\u003eVisone, R.\u003cem\u003e et al.\u003c/em\u003e Enhancing all-in-one bioreactors by combining interstitial perfusion, electrical stimulation, on-line monitoring and testing within a single chamber for cardiac constructs. \u003cem\u003eScientific Reports\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, (2018).\u003c/li\u003e\n\u003cli\u003eZoneff, E.\u003cem\u003e et al.\u003c/em\u003e Controlled oxygen delivery to power tissue regeneration. \u003cem\u003eNature communications\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, (2024).\u003c/li\u003e\n\u003cli\u003eEshmuminov, D.\u003cem\u003e et al.\u003c/em\u003e An integrated perfusion machine preserves injured human livers for 1 week. \u003cem\u003eNat Biotechnol\u003c/em\u003e \u003cstrong\u003e38\u003c/strong\u003e, (2020).\u003c/li\u003e\n\u003cli\u003eWhelan, J.\u003cem\u003e et al.\u003c/em\u003e In situ Raman spectroscopy for simultaneous monitoring of multiple process parameters in mammalian cell culture bioreactors. \u003cem\u003eBiotechnol Prog\u003c/em\u003e \u003cstrong\u003e28\u003c/strong\u003e, (2012).\u003c/li\u003e\n\u003cli\u003eStepula, E.\u003cem\u003e et al.\u003c/em\u003e Label-free 3D molecular imaging of living tissues using Raman spectral projection tomography. \u003cem\u003eNature communications\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, (2024).\u003c/li\u003e\n\u003cli\u003eClegg, J.\u003cem\u003e et al.\u003c/em\u003e Tailoring metabolic activity assays for tumour-engineered 3D models. \u003cem\u003eBiomaterials Advances\u003c/em\u003e \u003cstrong\u003e167\u003c/strong\u003e, (2025).\u003c/li\u003e\n\u003cli\u003eMartin, A. B.\u003cem\u003e et al.\u003c/em\u003e \u003cem\u003eHuman Anatomy and Physiology I\u003c/em\u003e. (LOUIS: The Louisiana Library Network, 2024).\u003c/li\u003e\n\u003cli\u003eKotecha, M. in \u003cem\u003eMagnetic Resonance Imaging in Tissue Engineering\u003c/em\u003e 21-47 (2017).\u003c/li\u003e\n\u003cli\u003eGil, C. J.\u003cem\u003e et al.\u003c/em\u003e Leveraging 3D Bioprinting and Photon-Counting Computed Tomography to Enable Noninvasive Quantitative Tracking of Multifunctional Tissue Engineered Constructs. \u003cem\u003eAdv Healthc Mater\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, (2023).\u003c/li\u003e\n\u003cli\u003eLiang, X.\u003cem\u003e et al.\u003c/em\u003e Imaging engineered tissues using structural and functional optical coherence tomography. \u003cem\u003eJournal of biophotonics\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, (2009).\u003c/li\u003e\n\u003cli\u003eTan, W.\u003cem\u003e et al.\u003c/em\u003e Optical coherence tomography of cell dynamics in three-dimensional tissue models. \u003cem\u003eOptics Express\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, (2006).\u003c/li\u003e\n\u003cli\u003eBland, E.\u003cem\u003e et al.\u003c/em\u003e Overcoming hypoxia to improve tissue-engineering approaches to regenerative medicine. \u003cem\u003eJ Tissue Eng Regen M\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, (2013).\u003c/li\u003e\n\u003cli\u003eEhrbar, M.\u003cem\u003e et al.\u003c/em\u003e Editorial: Advancing vascularized tissue models through biomaterials and biofabrication. \u003cem\u003eFront Bioeng Biotechnol\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, (2024).\u003c/li\u003e\n\u003cli\u003ePlace, T. L.\u003cem\u003e et al.\u003c/em\u003e Limitations of oxygen delivery to cells in culture: An underappreciated problem in basic and translational research. \u003cem\u003eFree Radical Biology and Medicine\u003c/em\u003e \u003cstrong\u003e113\u003c/strong\u003e, (2017).\u003c/li\u003e\n\u003cli\u003eWagner, B. A.\u003cem\u003e et al.\u003c/em\u003e The rate of oxygen utilization by cells. \u003cem\u003eFree Radic Biol Med\u003c/em\u003e \u003cstrong\u003e51\u003c/strong\u003e, (2011).\u003c/li\u003e\n\u003cli\u003eMagliaro, C.\u003cem\u003e et al.\u003c/em\u003e Oxygen Consumption Characteristics in 3D Constructs Depend on Cell Density. \u003cem\u003eFrontiers in Bioengineering and Biotechnology\u003c/em\u003e \u003cstrong\u003eVolume 7 - 2019\u003c/strong\u003e, (2019).\u003c/li\u003e\n\u003cli\u003eChastagnier, L.\u003cem\u003e et al.\u003c/em\u003e Deciphering dermal fibroblast behavior in 3D bioprinted dermis constructs. \u003cem\u003eBioprinting\u003c/em\u003e \u003cstrong\u003e32\u003c/strong\u003e, (2023).\u003c/li\u003e\n\u003cli\u003eHarcum, S. W.\u003cem\u003e et al.\u003c/em\u003e PID controls: the forgotten bioprocess parameters. \u003cem\u003eDiscover Chemical Engineering\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, (2022).\u003c/li\u003e\n\u003cli\u003eJones, J.\u003cem\u003e et al.\u003c/em\u003e Improved control strategies for the environment within cell culture bioreactors. \u003cem\u003eFood and Bioproducts Processing\u003c/em\u003e \u003cstrong\u003e138\u003c/strong\u003e, (2023).\u003c/li\u003e\n\u003cli\u003eChen, Y.\u003cem\u003e et al.\u003c/em\u003e Characterization and optimization of cell seeding in scaffolds by factorial design: quality by design approach for skeletal tissue engineering. \u003cem\u003eTissue Eng Part C Methods\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, (2011).\u003c/li\u003e\n\u003cli\u003eMartinez-Marquez, D.\u003cem\u003e et al.\u003c/em\u003e Application of quality by design for 3D printed bone prostheses and scaffolds. \u003cem\u003ePLOS ONE\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, (2018).\u003c/li\u003e\n\u003cli\u003ePapantoniou, I.\u003cem\u003e et al.\u003c/em\u003e Process Quality Engineering for Bioreactor-Driven Manufacturing of Tissue-Engineered Constructs for Bone Regeneration. \u003cem\u003eTissue Engineering Part C: Methods\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e, (2012).\u003c/li\u003e\n\u003cli\u003eBoubriak, O. A.\u003cem\u003e et al.\u003c/em\u003e Monitoring of metabolite gradients in tissue-engineered constructs. \u003cem\u003eJournal of the Royal Society, Interface\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, (2006).\u003c/li\u003e\n\u003cli\u003eLagarde, D.\u003cem\u003e et al.\u003c/em\u003e Lactate Fluxes and Plasticity of Adipose Tissues: A Redox Perspective. \u003cem\u003eFrontiers in Physiology\u003c/em\u003e \u003cstrong\u003eVolume 12 - 2021\u003c/strong\u003e, (2021).\u003c/li\u003e\n\u003cli\u003eYao, S.\u003cem\u003e et al.\u003c/em\u003e Role of lactate and lactate metabolism in liver diseases (Review). \u003cem\u003eInt J Mol Med\u003c/em\u003e \u003cstrong\u003e54\u003c/strong\u003e, (2024).\u003c/li\u003e\n\u003cli\u003eLi, X.\u003cem\u003e et al.\u003c/em\u003e Lactate metabolism in human health and disease. \u003cem\u003eSignal Transduction and Targeted Therapy\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, (2022).\u003c/li\u003e\n\u003cli\u003eBhatia, H.\u003cem\u003e et al.\u003c/em\u003e In-line monitoring of amino acids in mammalian cell cultures using raman spectroscopy and multivariate chemometrics models. \u003cem\u003eEng Life Sci\u003c/em\u003e \u003cstrong\u003e18\u003c/strong\u003e, (2017).\u003c/li\u003e\n\u003cli\u003eEsmonde-White, K. A.\u003cem\u003e et al.\u003c/em\u003e The role of Raman spectroscopy in biopharmaceuticals from development to manufacturing. \u003cem\u003eAnal Bioanal Chem\u003c/em\u003e \u003cstrong\u003e414\u003c/strong\u003e, (2022).\u003c/li\u003e\n\u003cli\u003eNtziouni, A.\u003cem\u003e et al.\u003c/em\u003e Review of Existing Standards, Guides, and Practices for Raman Spectroscopy. \u003cem\u003eAppl Spectrosc\u003c/em\u003e \u003cstrong\u003e76\u003c/strong\u003e, (2022).\u003c/li\u003e\n\u003cli\u003eUrciuolo, F.\u003cem\u003e et al.\u003c/em\u003e In vitro strategies for mimicking dynamic cell\u0026ndash;ECM reciprocity in 3D culture models. \u003cem\u003eFrontiers in Bioengineering and Biotechnology\u003c/em\u003e \u003cstrong\u003eVolume 11 - 2023\u003c/strong\u003e, (2023).\u003c/li\u003e\n\u003cli\u003eYi, J.-H.\u003cem\u003e et al.\u003c/em\u003e Remodeling of soft tissues due to cell activity. \u003cem\u003ePAMM\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, (2014).\u003c/li\u003e\n\u003cli\u003eAndilla, J.\u003cem\u003e et al.\u003c/em\u003e Imaging tissue-mimic with light sheet microscopy: A comparative guideline. \u003cem\u003eScientific Reports\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, (2017).\u003c/li\u003e\n\u003cli\u003eMatcher, S. J. Practical aspects of OCT imaging in tissue engineering. \u003cem\u003eMethods Mol Biol\u003c/em\u003e \u003cstrong\u003e695\u003c/strong\u003e, (2011).\u003c/li\u003e\n\u003cli\u003eHahn, A.\u003cem\u003e et al.\u003c/em\u003e in \u003cem\u003eTissue Engineering Using Ceramics and Polymers (Third Edition)\u003c/em\u003e (eds Aldo R. Boccaccini, Peter X. Ma, \u0026amp; Liliana Liverani) 281-343 (Woodhead Publishing, 2022).\u003c/li\u003e\n\u003cli\u003eXu, H.\u003cem\u003e et al.\u003c/em\u003e Monitoring tissue engineering using magnetic resonance imaging. \u003cem\u003eJ Biosci Bioeng\u003c/em\u003e \u003cstrong\u003e106\u003c/strong\u003e, (2008).\u003c/li\u003e\n\u003cli\u003eGuida, L.\u003cem\u003e et al.\u003c/em\u003e Advancements in high-resolution 3D bioprinting: Exploring technological trends, bioinks and achieved resolutions. \u003cem\u003eBioprinting\u003c/em\u003e \u003cstrong\u003e44\u003c/strong\u003e, (2024).\u003c/li\u003e\n\u003cli\u003eRicke, A.\u003cem\u003e et al.\u003c/em\u003e Magnetic Resonance Velocimetry for porous media: sources and reduction of measurement errors for improved accuracy. \u003cem\u003eExperiments in Fluids\u003c/em\u003e \u003cstrong\u003e64\u003c/strong\u003e, (2023).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7216794/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7216794/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAs the field of tissue engineering advances toward clinically viable, large-scale biofabrication, there is an urgent need for non-invasive, real-time monitoring tools capable of assessing the dynamic maturation of 3D bioprinted tissues. This study presents a modular analytical framework combining physicochemical, metabolic, morphological, and perfusion monitoring strategies tailored to volumetric engineered tissues. A perfused cultivation platform was developed for 10.8 cm\u0026sup3; bioprinted fibroblast tissues, enabling fine regulation of pH, temperature, and oxygen. Enhanced oxygen control was achieved through dual-gas PID regulation, reducing deviation from 128\u0026ndash;22%. Metabolic activity was monitored via online Raman spectroscopy, allowing real-time lactic acid quantification with a prediction precision error of 0.103 g.L⁻\u0026sup1;, despite low secretion levels typical of adherent cells. Morphological evolution was tracked using 7 Tesla MRI, revealing high fidelity to initial designs (87.6% within 1 mm deviation) and providing longitudinal insights into tissue remodeling without labeling or sectioning. Perfusion was evaluated through computational fluid dynamics (CFD) simulations and MRI velocimetry, confirming flow heterogeneity and validating internal fluid distribution. These combined approaches demonstrate the feasibility of a closed-loop, feedback-driven biomanufacturing process that aligns with quality-by-design principles and regulatory expectations for advanced therapy medicinal products (ATMPs). The integration of established tools from pharmaceutical and clinical fields into tissue engineering workflows marks a critical step toward scalable, standardized, and adaptive biofabrication processes capable of supporting the next generation of functional tissue substitutes.\u003c/p\u003e","manuscriptTitle":"Exploring multiple monitoring modalities for large-scale 3D tissue cultivation in bioreactor","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-22 11:43:40","doi":"10.21203/rs.3.rs-7216794/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-10-08T08:50:06+00:00","index":"","fulltext":""},{"type":"reviewerAgreed","content":"81358417598781931953437720717206242723","date":"2025-09-18T09:03:33+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-09-15T12:38:29+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"109775566284153859791277467883146916637","date":"2025-09-14T09:05:48+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-09-08T02:38:30+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"242062006544254371229030490666354612868","date":"2025-08-17T19:14:38+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-08-14T11:19:20+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-08-14T11:07:23+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-08-12T05:45:29+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-08-08T11:43:32+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-08-08T11:40:17+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"4e2c355d-7577-4787-a975-9bb15be5d3e8","owner":[],"postedDate":"August 22nd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":53267968,"name":"Biological sciences/Biological techniques"},{"id":53267969,"name":"Biological sciences/Biotechnology"},{"id":53267970,"name":"Physical sciences/Engineering"}],"tags":[],"updatedAt":"2026-01-12T16:13:02+00:00","versionOfRecord":{"articleIdentity":"rs-7216794","link":"https://doi.org/10.1038/s41598-025-29257-y","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2026-01-07 15:58:29","publishedOnDateReadable":"January 7th, 2026"},"versionCreatedAt":"2025-08-22 11:43:40","video":"","vorDoi":"10.1038/s41598-025-29257-y","vorDoiUrl":"https://doi.org/10.1038/s41598-025-29257-y","workflowStages":[]},"version":"v1","identity":"rs-7216794","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7216794","identity":"rs-7216794","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}