Abstract
Cell Organoid Engineering and Cell Therapy
Accelerated and Improved Differentiation of Retinal Organoids from Pluripotent Stem Cells in Rotating-Wall Vessel Bioreactors
Three-dimensional (3D) organoid cultures have gained in popularity in recent times for investigating morphogenesis, disease pathology, and therapeutic development. The retina is a complex tissue responsible for vision and involves organizing multiple neuronal subtypes to form intricate synaptic circuits. A 3D organoid culture involving pluripotent (mouse, human) and matrigel has led to the generation of optic cup structures reminiscent of in vivo retina development using hypoxia (<5% O2) and defined media culture. However, strong evidence suggests that 3D organoids are diffusion limited without a reliable nutrient supply (e.g., vascular system) for nutrient and waste exchange.
Rotary wall vessel bioreactor culture has been used to maintain cell aggregates and tissue constructs in stationary suspension by balancing the settling aggregates with the motion of the medium. This keeps the cells/tissue constructs in simulated microgravity, providing efficient mass transfer while reducing shear stresses. It also facilitates the culture of shear-sensitive cell types.
The authors demonstrate that the rotary bioreactor promotes retinal organoid growth and differentiation. Compared with static cultures, the rotary bioreactor culture recapitulates in vivo mouse retina development to a greater extent.
The retina tissue development protocol lasts approximately 25 d (10 d in static matrigel culture) with stages to successively mature the pluripotent stem cells. Because of the growth of the cell aggregates, the rotation speed is increased from days 11 through 25 to maintain the stationary aggregate suspension. Two conditions of static culture (dissected and intact neural retina [NR]) are compared with rotary cultures. Even though dissected NR facilitates improved growth, the rotary culture growth surpasses the two other conditions. With culture time, the phosphohistone 3 biomarker (indicating mitotic cells) reduces in expression for all cultures, especially for rotary culture.
Next, the differentiation progress is analyzed in the different cultures. Retinal ganglion cells, horizontal, amacrine, and S-cone–positive cells are promoted by rotary cultures. When later-stage biomarkers are examined, rhodopsin and glutamine synthetase (rod photoreceptors and Müller glia, respectively), the rotary cultures are found to significantly enhance and accelerate their expression compared with static cultures. Increased bipolar cells, photoreceptors, and synapse formation are observed to a greater extent in rotary compared with static cultures.
To objectively assess gene expression trends, RNA sequencing is performed to compare the conditions to in vivo retina development. Although the three conditions exhibit high similarity, rotary culture exhibits accelerated differentiation. For example, day 25 of rotary cultures are more comparable to day 28 static (dissected) and day 32 static (intact).
Thus, the authors show that rotary cultures are capable of promoting in vivo–like characteristics in NR organoids. Rotary cultures are able to mature the NR organoids to a developmental stage equivalent to P6 mice (day 25, rotary). Evidence also suggests that the P6–P10 transition requires additional cues not present in organoid culture. Further developments in this platform may involve soluble cues, retinal pigment epithelium (co-cultures), and even biomaterial matrices. (Di Stefano, T.; et al. Stem Cell Rep.
Organ-Specific Metastases Obtained by Culturing Colorectal Cancer Cells on Tissue-Specific Decellularized Scaffolds
To date, there is a general lack of understanding of the interaction between tissue-specific microenvironments and metastatic cells. Conventional in vitro systems (e.g., 3D, collagen, matrigel) cultures differ largely from tissue-specific microenvironments. Another option, genetically engineered animal models that develop metastases, is costly and difficult to use.
An alternative to studying site-specific interactions of cancer cells involves investigating their interactions with extracellular matrices from organs. Using perfusion to strip organ tissue of cells, the authors of this report use the remaining support substrate (AKA biomatrix scaffolds) to study metastatic colorectal cancer of the liver and lung as a proof of concept. The biomatrix scaffold is then coated on culture dishes before seeding colorectal cancer cells.
The decellularization process retained >98% of matrix components and preserved physiological levels of matrix-bound growth factors and cytokines. Enzyme-linked immunosorbent assay and mass spectrometry measurements show that growth factors and cytokines are largely retained following decellularization (93%). From cell culture, several colorectal carcinoma (CRC) cell cultures generate tumor spheroids of millimeter dimension, whereas those cultured on plastic, collagen, and matrigel fail to do so. Excitingly, this suggests that the biomatrices promote and support the growth of in vivo–like metastasis tissue in vitro.
Further exploration of the generated tissue shows morphological features reminiscent of in vivo metastasis. These include (1) signet ring cells, (2) bizarre mitotic figures, (3) necrotic debris, (4) pleomorphic cell size and shape, and (5) multinucleated cells. Signet ring cells, found in animal metastases, have not been previously reported in other ex vivo model systems to date (e.g., matrigel cultures). Microarray analysis of cells from animal liver metastases (in vivo) and those obtained from liver biomatrices by these authors show high levels of similarity, unlike matrigel and plastic cultures. Subsequently, CRC cells from the different matrices are implanted into mice for in vivo evaluation. Once again, the lung and liver biomatrices promote the greatest in vivo proliferation. The biomatrix-cultivated CRCs also show greater invasive potential in basement membrane models.
Finally, the biomatrix-cultivated CRCs allow the authors to investigate responses to different therapeutic regimes. These include 5-fluorouracil (5FU), irinotecan alone, irinotecan + 5FU, oxaliplatin alone, and oxaliplatin + 5FU and radiotherapy. Crucially, they discover that liver metastases are more robust to chemotherapy (compared with other in vitro culture methods). Differential radiotherapy responses are observed for cells cultured with different substrates. Biomatrix cultures are shown to be an efficient method to recapitulate in vivo metastases in culture. An intriguing possibility from this proof of concept is the development of high-throughput screening assays to identify therapeutics to treat metastases in an organ-specific way. (Tian, X.; et al. Nat. Biomed. Eng.
An Integrated Miniature Bioprocessing for Personalized Human-Induced Pluripotent Stem Cell Expansion and Differentiation into Neural Stem Cells
Although induced pluripotent stem cells (iPSCs) have emerged as very favorable candidates for personalized cell therapy, their effective translation is often limited by conventional cell culture infrastructure. Cell biomanufacturing involves many steps, including culture, reprogramming, clone selection, expansion, and characterization, before further expansion and differentiation into desired cell types can be achieved. The generated cells are then purified and characterized before transplantation. Further logistical impediments include a requirement for skilled operators, compliance with current Good Manufacturing Process, transportation, tracking, and reporting.
To circumvent obstacles in bioprocessing, the authors describe an integrated mini-bioprocess for developing neural stem cells (NSCs) from human iPSCs. Human iPSCs are first expanded in 3D thermoreversible hydrogels before NSC differentiation. Within this integrated bioprocess, undifferentiated iPSCs are depleted, concentrated, and transported within closed 15-mL conical tube reactors.
iPSCs are encapsulated in thermoreversible gels and grown as 3D cultures, and H9 human embryonic stem cells (ESCs) are grown alongside as controls. Below 4 °C, the gel is a liquid polymer solution, and it then forms an elastic hydrogel above room temperature. Grown in a single-cell state, iPSCs undergo clonal expansion into uniform spheroids with approximately 20-fold expansion within 5 d. Neural differentiation can be triggered by inhibiting SMAD signaling in defined E8 media. This is performed by removing basic fibroblast growth factor and transforming growth factor–β and supplying LDN193189 and SB431542 to inhibit SMAD signaling. Initially, even though >95% of cells express OCT4 and NANOG, less than 2% are positive for pluripotent markers after 7 d of differentiation. Instead, 95% of cells on day 7 express NESTIN and PAX6–NSC biomarkers. Throughout, monolayer (2D) and 3D hydrogel cultures display no significant differences in biomarker expression. H9 ESCs also perform similarly to iPSCs. Further culture led to TUJ1+ and TH+ neuronal formation. This demonstrates how the differentiation process can generate more specialized brain cells from pluripotent progenitors.
Finally, the authors develop a single-reactor bioprocess, integrating iPSC growth and differentiation. Reagents are removed through a sterile syringe placed through the septum cap. Magnetic beads coated with SSEA4 antibodies are then added to isolate undifferentiated iPSC using magnetic separation. Purified cells are then transferred into a fresh tube and transported for surgical transplantation in rats. These are then successfully integrated with the rat brain and identified through dual human nuclear antigen and TUJ1+ biomarker expression. To summarize, the miniaturized bioprocess could benefit cell therapy by performing culture in an efficient and cost-effective way that integrates several actions (e.g., encapsulation, expansion, differentiation, purification, and transfer) within a single receptacle. (Lin, H.; et al. Sci. Rep.
Unraveling the Inconsistencies of Cardiac Differentiation Efficiency Induced by the GSK3β Inhibitor CHIR99021 in Human Pluripotent Stem Cells
The signaling pathway GSK3β is critical to control fate decisions of stem cells. GSK3β phosphorylates β-catenin, leading to its degradation. Cardiac differentiation has been demonstrated in pluripotent stem cells (PSCs) by GSK3β inhibition via the small-molecule inhibitor CHIR99021 (CHIR). However, this protocol requires culture optimization and often results in heterogeneous outcomes. This led the authors of this report to investigate CHIR induction in PSCs and how it facilitates mesoderm formation and cardiac differentiation.
Because CHIR is a kinase inhibitor of GSK3α and GSK3β, it potentially has off-target effects on cytotoxicity, cell growth, and the cell cycle. The first investigation involves examining culture confluency (low: <50%, mid: 70%, and high: >90%) on embryoid body (EB) formation (cell aggregation to coax PSCs to differentiate) and monolayer (2D) culture. Although CHIR is found to be essential for EB formation, it induces cell death and growth dose dependently. In cells with low (<41%) S/G2/M cell-cycle phase, CHIR is significantly cytotoxic, but in high (>42%) S/G2/M cell-cycle phase, CHIR increases cell mass. The act of cardiac differentiation itself causes transient cell death. CHIR induces cell cycle gene cyclin D1 expression, increasing cells in the S/G2/M cell-cycle phase. This leads to increased growth in cultures with <70% confluency.
Using ESCs with NKX2.5 reporter genes, they show how lower cell densities favor cardiac differentiation. At more than 90% confluency, cardiac commitment drops from 50% to 15%. Further analysis illustrates how identifying cell lines with high levels of initial pluripotency (Nanog and OCT4 biomarkers) can identify whether efficient cardiac differentiation occurs. CHIR induction is also shown to be regulated by T-cell factors (TCFs) that are activated within 4 to 24 h in a dose- and culture condition–dependent manner.
The role of CHIR activation in cardiac differentiation spurs primitive streak and mesoderm development in different culture conditions (2D, EBs, and microcarrier cultures). Although mesoderm differentiation is fairly robust (to differences in culture conditions), cardiac maturation is much more sensitive, being affected by various cell density and cell-cycle conditions. When CHIR is depleted, TCF expression similarly decreases. On the other hand, external Wnt inhibition (the IWR-1 chemical) prolongs TCF expression, supporting mesoderm and early cardiac specialization.
In summary, Oh and colleagues are able to demonstrate how GSK inhibition connects cell-cycle and cell-fate decisions to direct cardiac development. These factors also determine proliferation and cell death decisions. By elaborating on cell-cycle interplay with CHIR induction, it accounts for many instances of seemingly paradoxical behavior of CHIR-induced cardiac differentiation in pluripotent stem cells. Repeating these studies with other GSK3β kinase inhibitor drugs (apart from CHIR) would make these findings more applicable to GSK3β-triggered cardiac differentiation. Another point worth highlighting is that the cell-cycle profile emerges as a critical quality control factor in selecting candidate PSCs with high differentiation efficiency. (Laco, F.; et al. Stem Cell Rep.
Efficient Ex Vivo Engineering and Expansion of Highly Purified Human Hematopoietic Stem and Progenitor Cell Populations for Gene Therapy
Lentiviral (LV)–based gene therapy of ex vivo engineered, autologous CD34+ hematopoietic stem cells (HSCs) has been used for correcting genetic disorders with promising results in clinical trials. Although the perception is that its safety levels are still undesirable, no adverse insertional mutagenesis events (semi-randomly integrating LVs) have been reported to date, despite >150 patients receiving treatment. A bottleneck in the cell therapy process is that HSCs progressively lose engraftment potential in culture when they undergo cell-cycle recruitment and lose adhesion molecules, which impedes niche homing. This further encourages lineage commitment and differentiation, losing highly potent stem cells with high engraftment potential.
CD34+ hematopoietic stem progenitor cells often comprise a heterogeneous mixture of cells with a mixture of various stages of potency/lineage commitment. Some studies even suggest that among these, only a mere ~0.01% of infused cells (a small subpopulation) contribute to long-term hematopoiesis.
Given these considerations, the authors develop a comprehensive strategy to examine HSC populations by uncoupling long- and short-term hematopoietic reconstitution and implement ex vivo culture conditions to best preserve biological properties with chemical compounds and support undifferentiated HSC expansion. Their goal is to maximize the persistence of HSC engraftment by identifying, then optimizing, different factors that influence HSC culture.
To identify which subpopulation of CD34 cells contributes to hematopoiesis, the authors tag different subpopulations using different fluorescent reporters based on their CD38 expressions. The CD34stem population (lowest 10% CD38 expression) exhibited engraftment up to 24 wk, albeit to lower levels compared with the CD34total (containing more nonstem progenitor cells) population. Thereafter, the authors use the CD90 biomarker to further discriminate the population based on the gradient of its expression (another four fluorescent colors). This secondary purification reveals that one-third of engrafted cells originate from a CD90-bearing subpopulation. Through further analysis comparing LV-transfected cells and the whole CD34 population, the long-term repopulating capacity is deemed to significantly diminish during the 2 d of ex vivo culture. Thus, this provides a motivation to reduce culture times to improve cell transplant potency.
Further modifications to the protocol include shortening culture time and eliminating interleukin-3 from the culture medium, which increases the engraftment population to 80% of the cells. Further experimentation shows that the highly potent HSC populations (CD34+CD38–) unexpectedly transduce twofold higher amounts of LV compared with the more committed CD34total population. Applying the small-molecule factor –16,16-dimethyl prostaglandin E2 as an antiapoptotic agent further increases gene-integration ability. This specifically enhances the viral copy number in the long-term repopulation of cells, which leads to differences maintained even 18 to 20 wk after transplantation.
Because the authors show that the short-term repopulating cell fraction is rapidly exhausted in ex vivo culture, they employ 2 small molecules, StemRegenin 1 (an arylhydrocarbon receptor antagonist) and UM171 (a pyrimidoindole derivative). These enable the authors to boost cell numbers by performing ex vivo expansion for a period of 8 d while still maintaining high engraftment potential due to the further optimized conditions.
This study shows how examination of the various factors involving HSC ex vivo expansion and gene transduction can be optimized in a stepwise fashion. Further optimization will enhance HSC gene therapy/correction and improve the likelihood of genetic blood disorder therapy in eventual clinical trials. (Zonari, E.; et al. Stem Cell Rep.
Nano- and Microtechnologies
A Ribonucleoprotein Octamer for Targeted siRNA Delivery
Delivery of nucleic acids (NAs) in vivo is limited by poor loading efficiency and inherent charge issues. For example, positively charged carriers allow better loading but cause major protein aggregation in ECM and blood. The authors of this report propose delivering siRNA using a novel system containing an 8-arm polyethylene glycol scaffold carrier using H2E peptides for endosome rupture and escape and connecting dsRNA binding domain through azide click chemistry. These carry siRNA targeting polo-like kinase 1, which induce apoptosis that conjugates with small-molecule 2-[3-(1,3-dicarboxypropyl) ureido]pentanedioic acid (DUPA) to recognize prostate-specific membrane antigen (PSMA). The NA delivery strategy was termed RNP8.
Using various mass spectrometry, gel electroporation, transmission electron microscope, and dynamic light-scattering methods, the authors validate the generation of ~32 nm nanoparticles. Incubation in mouse serum shows sufficient protein-siRNA stability with a half-life exceeding 18 h. To test this in cells, the RNP8 nanoparticles show PSMA-specific uptake in different prostate cancer cell lines, including PC3 (negative PSMA), 22RV1 (positive PSMA), and LNCaP cells (highly positive PSMA). Treatment with the PSMA inhibitor 2-(phosphonomethyl)-pentanedioic acid further demonstrated RNP8 specificity when applied to LNCaP and reversing its high uptake levels.
The authors then observe that the structure starts to disassemble 3 h following uptake, without signal colocalization between carrier, cargo, and endosomes. In cell culture conditions, RNP8 achieves 60% to 74% silencing (100–300 nM) compared with ~90% by lipofectamine at 100 nM concentration. Despite their superior performance, lipofectamine and other cationic carriers are unsuitable for in vivo usage. RNP8 caused G2-M arrest, which leads to reduced proliferation and apoptosis.
In vivo distribution is next assessed through intravenous delivery. Biodistribution is significantly improved over RNP8 (without DUPA) or naked siRNA/DUPA. Its silencing efficacy is also assessed and applied once every 3 d with 11 doses applied in total. RNP8’s specificity for PSMA is shown in the retardation of LNCaP tumor growth but not those of PC3. Further analysis with the terminal deoxynucleotidyl transferase dUTP nick-end labeling assay confirms that significant numbers of apoptotic cells are found within the tumor.
Critical to in vivo efficacy is stealthing the payload from immune responses. The authors conclude with a study on human PBMC (72 h incubation) and in mice (2 h postinjection). These showcase the advantages of the RNP8 platform as cytokine production is found to be minimal compared with siRNA-DUPA. In contrast to the benign effects of RNP8 on immunity, cationic polyethyleneimine-conjugated siRNA elicits a strong immune response. (Tai W.; et al. Nat. Biomed. Eng.
Diagnosis of Sepsis from a Drop of Blood by Measurement of Spontaneous Neutrophil Motility in a Microfluidic Assay
Accurate sepsis diagnosis remains elusive and misdiagnosed in ~30% of patients. This leads to unnecessary antibiotic usage that indirectly promotes antibiotic resistance. A number of systemic biomarkers have been proposed, including C-reactive protein, procalcitonin, and interleukin-6. Neutrophil cell biomarkers such as CD64 have also been explored, but they provide limited diagnostic value. Microbiology cultures can diagnose sepsis, but they require at least 2 to 3 d, which is unamenable for early diagnosis and treatment.
Neutrophils derived from patients with sepsis have altered function compared with uninfected cells. Because of this, the author’s previous work showed that isolated neutrophils in microfluidic channels provides a sepsis prediction value of 80% sensitivity and 77% specificity in patients with major burns.
Because of the sensitivity of neutrophils to factors released from inflammation and infection responses, the authors hypothesize that measuring neutrophil motility with whole-blood samples may generate more significant changes compared with isolated neutrophils used in previous work. In this article, Daniel Irima and colleagues engineer microfluidic devices to gauge neutrophil activity to diagnose sepsis. Cell movement (motility) through mazes of channels is tracked to diagnose their phenotype. Interactions between neutrophils and plasma are critical for their respective functional behavior. By tracking this behavior with image analysis, septic diagnosis predictions are better than 98%. This device contains a 1 µL sample volume of diluted blood, a red blood cell filter (which also traps other leukocytes that are too large and undeformable), as well as migration channels and a maze. Following deposition of dilute whole-blood samples, neutrophils are found to consist of >96% total cells.
From clinical blood samples, neutrophil motility characteristics are observed using unbiased machine learning followed by further supervised machine-learning analysis. Ultimately, five parameters are found to confer the greatest sepsis prediction accuracy. These were neutrophil count, number of oscillations made by the neutrophils, number of cells pausing, reverse migration out of the device, and the average distance of migration. These parameters are scored and found to generate an area under the receiver operator characteristic curve (AUC) of 0.98 for patients with and without sepsis. Correspondingly, this method achieved 96.8% sensitivity and 97.6% diagnostic specificity.
This assay does not use chemoattractants and relies on spontaneous motility that increases its robustness. It requires 4 h of time-lapse microscopy and 2.5 h of image processing and analysis. Further studies also find that blood testing is optimal within the first 3 h of collection, as neutrophil activity diminishes with age. Thus, minimizing the time between collection and assay benefits diagnostic precision. Whole-blood assays are found to magnify functional behavior compared with isolated neutrophils. This abnormal septic neutrophil phenotype is found to be a result of cell-intrinsic factors and also extracellular factors borne in the plasma.
The assay is then carried out on a second, independent cohort of 19 (10 with sepsis and 9 without) patients that was double-blinded and performed as a prospective case-control study. The second independent cohort generates similar results to the first validation cohort, resulting in an overall AUC of 0.99, with 97% sensitivity and 98% specificity for the total of 42 patients. The contrasting performance between diseased and healthy neutrophils truly demonstrates their critical role in diagnosing sepsis.
In summary, this assay is logistically simple (foregoing any cell isolation steps) and easy to use (prepare and load blood samples), using an automated image analysis program that generates a sepsis score. Obtaining larger and more diverse cohorts of patients improves this assay’s method, analysis, and threshold values and allows sepsis to be effectively diagnosed. (Ellet, F.; et al. Nat. Biomed. Eng.
Cardiac Recovery via Extended Cell-Free Delivery of Extracellular Vesicles Secreted by Cardiomyocytes Derived from Induced Pluripotent Stem Cells
Currently, standard heart disease therapy cannot regenerate the injured myocardiac tissue and does not alleviate demand for heart transplants. Cell therapy has shown modest improvements, but its efficacy is limited by poor retention and concerns for tumorigenicity. A promising approach in recent times is extracellular vesicles (EVs) that attenuate ischemic injury; however, it suffers from short-lived retention times (<3 h).
To circumvent such limitations, the authors of this report propose using hydrogel patches to encapsulate and gradually release EVs from cardiomyocytes (iCMs) derived from iPSCs. Characterization of EVs isolated from iCMs shows EV biomarkers such as tumor susceptibility gene 101 and the absence of cell biomarkers that confirms their EV status. The iCM-EVs are also readily uptaken by cardiac and endothelial cells, evidenced by internalization. Using in vitro culture, the authors show that the iCM-EVs confer more therapeutic potential in hypoxic cultures compared with iPSC-EVs derived from parental cells. Similarly, the iCM-EVs protect vascular cells in culture by preserving vascular networks, which demonstrates their therapeutic potential. miRNA sequencing reveals that iCM-EVs are highly abundant in miR-1 and miR133a and are significantly known to affect cardiac function. Gene ontology analysis also reveals that these miRNAs are involved in cardiac biology. Because of miRNA’s inhibitory function, this suggests that they positively regulate cardiac health by suppressing hypertrophy.
The authors generate a therapeutic patch by encapsulating 3 × 1010 iCM-EVs into a 7 mm collagen gel foam mesh. These EVs are completely released over a period of 21 d, with similar kinetics in vivo (rat model). A myocardial infarction is created by ligation in the left anterior descending artery in athymic nude rats (to negate any immune reactions). Patches containing iCM-EVs, iPSC-EVs, and saline are then applied directly into the rat myocardium after the ligation. Further investigation shows that the EVs deliver their miRNA cargos from the myocardium patches.
By monitoring the ligated subjects with electrocardiograms, left anterior descending ligation generates arrhythmic events as expected. These events are significantly increased in the ligation group compared with the sham group, although this is suppressed when iPSC and iCM-EVs are introduced. When the rats are stressed with isoproterenol that causes ventricular arrhythmia, the EV patches significantly reduce this arrhythmic burden.
Using echocardiograms, the authors show that iCM-EV treatment reduces cardiac dilation and exhibits better ejection fraction compared with other groups (including iPSC-EVs). With the iCM-EVs, anterior wall motion is greater compared with the untreated group. This demonstrates the extent of the therapeutic EVs’ protection against declining myocardiac function. Further examination of fibrotic tissue also shows significantly lesser quantities in the iCM-EVs. This is found to be due to the treatment preventing cell apoptosis in hypoxic (<1% O2 conditions).
Crucially, this study demonstrates how hydrogel patches can be engineered to generate sustained release of EVs that possess therapeutic function following the induction of myocardial infarction. The treated rats demonstrate better heart function and fewer irregular events through the prevention of cell apoptosis, which contributes to less fibrotic tissue generation. There are also benefits from the cell-free mechanism, circumventing any limitations incurred by cell therapy. (Liu, B.; et al. Nat. Biomed. Eng.