Leveraging 3D Bioprinting and Photon-Counting Computed Tomography to Enable Noninvasive Quantitative Tracking of Multifunctional Tissue Engineered Constructs

Abstract

3D bioprinting is revolutionizing the fields of personalized and precision medicine by enabling the manufacturing of bioartificial implants that recapitulate the structural and functional characteristics of native tissues. However, the lack of quantitative and noninvasive techniques to longitudinally track the function of implants has hampered clinical applications of bioprinted scaffolds. In this study, multimaterial 3D bioprinting, engineered nanoparticles (NPs), and spectral photon-counting computed tomography (PCCT) technologies are integrated for the aim of developing a new precision medicine approach to custom-engineer scaffolds with traceability. Multiple CT-visible hydrogel-based bioinks, containing distinct molecular (iodine and gadolinium) and NP (iodine-loaded liposome, gold, methacrylated gold (AuMA), and Gd₂ O₃ ) contrast agents, are used to bioprint scaffolds with varying geometries at adequate fidelity levels. In vitro release studies, together with printing fidelity, mechanical, and biocompatibility tests identified AuMA and Gd₂ O₃ NPs as optimal reagents to track bioprinted constructs. Spectral PCCT imaging of scaffolds in vitro and subcutaneous implants in mice enabled noninvasive material discrimination and contrast agent quantification. Together, these results establish a novel theranostic platform with high precision, tunability, throughput, and reproducibility and open new prospects for a broad range of applications in the field of precision and personalized regenerative medicine.

Keywords: 3D bioprinting; bioinks; contrast agents; longitudinal quantitative imaging; nanoparticles; photon-counting computed tomography; tissue engineering scaffolds.

Figure 1.. Schematic illustration of the research workflow in this study.

A: Various computed tomography (CT) contrast agents were incorporated in gelatin methacrylate (GelMA), with and without polyethylene glycol dimethacrylate (PEGDMA), to create functionalized bioinks. These included molecular iodine (Iod) and gadolinium (Gd) contrast agents (i), Iod-loaded liposome nanocapsule (ii), and nanoparticle (NP) contrast agents, namely Gd₂O₃, gold (Au), and methacrylated gold (AuMA) NPs (iii). A one-step UV-photopolymerization of GelMA and AuMA NPs was used to form covalent linkages and prepare GelMA-Au NP hydrogels (iv). Adapted from [53]. B: CAD models of scaffolds with varying geometries were created and assigned to distinct CT contrast agent-laden bioinks. C: 3D bioprinting of CT-visible bioink formulations was conducted using either extrusion (Ex) (i) or digital light processing (DLP)-based (ii) techniques to create 3D constructs of varying geometry (iii). D: Printed constructs were crosslinked immediately post print under varying UV intensity for a 4-min duration (2 x 2 mins, flipping the construct at midpoint). E: A group of extrusion-printed constructs (simple disc geometry) were used to study contrast agent release under static (i) vs. dynamic (ii) conditions in PBS. At each time point, constructs were retrieved from the solution, embedded in agar, and stored at −80°C (to avoid further release) until micro-CT imaging. F: A separate group of printed constructs consisting of multiple distinct contrast agents in varying geometries were immediately embedded in agar, stored at −80°C, and imaged via micro-CT and photon-counting CT (PCCT) either in vitro (in agar) (i) or after subcutaneous implantation in the mouse torso (ii). G: CT imaging of various experimental groups were performed either via micro-CT (i) (for release study) or spectral PCCT (ii). H: Structural, mechanical, and biological characterization of bioinks and printed constructs were performed using micro, macro, and volumetric fidelity analyses (i), microindentation and rheological analyses (ii), and cell viability and growth assays (iii).

Figure 2.. Characterization of printing fidelity and mechanical properties of bioprinted CT-visible GelMA constructs.

A: The design of a two-layer model used to assess 2D micro-scale (strand-level) fidelity. B: Quantification of micro-scale fidelity parameters for various bioinks, including the strand diameter ratio ($r_d$), angle ratio ($r_{\alpha }$), uniformity ratio ($r_U$), and inter-strand area ratio ($r_A$) (n = 5 per group). The 12% GelMA-based bioink groups included the bare GelMA (control), GelMA containing gadolinium (Gd) and iodine (Iod) molecular contrast agents, and GelMA containing Gd₂O₃ and Iod-loaded liposome nanoparticles (NPs). C: Qualitative evaluation of printed strands consisting of various bioinks via bright-field imaging of strands at two magnifications. D-E: Quantitative (D) and qualitative (bright field images) (E) analysis of strand fidelity, conducted for various 20% GelMA-based bioink formulations (n = 5 per group), including pure GelMA (control), and those loaded with gold (Au) and methacrylated gold (AuMA) NPs. F: Macro-scale fidelity analysis was conducted for all bioink groups by measuring the diameter of a disc-shape construct and comparing with the CAD values (disc diameter ratio, $r_D$) (n = 4 per group). G: Microindentation test (i) was conducted on 12% (ii) and 20% (iii) GelMA-based bioink formulations to quantify the elastic modulus (E) (n = 6 per group). Scale bars in top rows in C and E represent 2 mm and in bottom rows represent 1 mm. *: P < 0.05, ***: P < 0.005, and ****: P < 0.001. n.s.: not significant.

Figure 3.. In vitro characterization of the retention of various molecular and nanoparticle (NP) contrast agents from 3D bioprinted GelMA constructs.

A: Standard curves of Gd chelate and Iod (molecular) contrast agents. B-E: In vitro release of Gd chelate from bioprints under (B-C) static and (D-E) dynamic (rocking) conditions in PBS for a 1-hr duration (n = 4 per group). C,E: Corresponding micro-CT reconstructions at t = 0 and 1 hr post-incubation. ‘No signal’ labels indicate that no detectable CT signal was obtained at indicated time points. F-I: In vitro release of Iod from bioprints under (F-G) static and (H-I) dynamic conditions in PBS for a 2-hr duration (n = 4 per group). G,I: Corresponding micro-CT reconstructions at t = 0 and 2 hrs post incubation. J: Standard curves of Gd₂O₃ and Au NP contrast agents. K-L: In vitro release of Gd₂O₃ NPs from bioprints under dynamic conditions in PBS for a 504-hr (3-week) duration (n = 4 per group) (K), and corresponding micro-CT reconstructions at t = 0 and 2 hrs (L). M-N: In vitro release of Iod from Iod-loaded liposome nanocapsules in the bioprints under dynamic conditions in PBS for a 2-hr duration (n = 4 per group) (M) and the corresponding micro-CT reconstructions (N) at t = 0 and 2 hrs. O-P: In vitro release of Au NPs from bioprints under dynamic conditions in PBS for a 504-hr (3-week) duration (n = 4 per group) (O) and the corresponding micro-CT reconstructions (P) at t = 0 and 2 hrs. Q-R: In vitro release of AuMA NPs from bioprints under dynamic conditions in PBS for a 504-hr (3-week) duration (n = 4 per group) (Q) with the corresponding micro-CT reconstructions (R) at t = 0 and 2 hrs. All scale bars represent 1 mm. *: P < 0.05, **: P < 0.01, ***: P < 0.005, and ****: P < 0.001.

Figure 4.. Spectral photon-counting computed tomography (PCCT) of bioprinted constructs of varying geometries, containing multiple contrast agents and their quantitative analysis.

Four different scaffold geometries were printed, including an (A) concentric circle (ConC), (B) rectangle in circle (RinC), (C) vascular tree in circle (VasC), and (D) Georgia Tech (GT) logo. For each scaffold geometry (A-D), row (i) shows the CAD model of the design with Gd₂O₃ NP and AuMA NP bioinks depicted in different regions. Row (ii) shows photos of extrusion-bioprinted constructs from side (left) and axial/top (right) views where Gd₂O₃ bioinks appear white and AuMA NP bioinks appear black. Row (iii) shows conventional CT image slices for each scaffold design, with X-ray attenuation in grayscale (Hounsfield unit, HU). Row (iv) shows the lowest-energy bin PCCT axial slice (top left) in grayscale (HU) along with axial (top right) and coronal (bottom left) slices of the color-coded Gd (green) and Au (red) decomposed images, as well as a 3D rendering (bottom right). Material map slices are windowed according to the color bars on the far right of this figure. E-F: Fidelity quantification based on the PCCT imaging of (E) ConC and (F) RinC designs, obtained by normalizing the dimensional measurements in PCCT images by those in the CAD file (left), or by those in the bioprinted construct (right) (n = 3 per group). Insets show representative PCCT images used for the quantifications. Scale bars in A-B and E-F show 2 mm and in C-D show 4 mm. All concentration values are in mM.

Figure 5.. Spectral photon-counting computed tomography (PCCT) imaging of subcutaneous implants in mice to track the location of bioprinted constructs and quantify the concentration of Gd₂O₃ and AuMA NP contrast agents in the scaffolds.

A: Wild type mice were bilaterally implanted post-mortem with a bioprinted disc-shape construct containing no contrast agent (control, right side of the chest, the blue arrows) and a bioprinted concentric circle (ConC) implant containing Gd₂O₃ and AuMA NP contrast agents (left side of the chest, the white arrows). B-D: Grayscale PCCT images from a single-energy bin, including a 3D rendered image (B), axial slice (C), and two oblique slices through the implanted constructs (top and bottom) (D). Images are shown in Hounsfield unit (HU) as displayed on the far-right side (−1000 to 1500 HU). E-G: Color-coded images of Gd (green) and Au (red) material basis maps, including a 3D rendered image (E), axial slice (F), and two oblique slices through the implanted constructs (G). Au and Gd maps are shown in mM concentration. Each map was windowed according to the color bars to the right of the decomposed images and then combined as a composite RGB image. Note that the bones appear yellow in the decomposed images. Blue and white arrows point to the control and ConC implants. Scale bars show 5 mm in all images.