This paper presents methods for growing cardiac myocytes with different shapes, which represent different pathologies, and sorting these adherent cardiac myocytes based on their morphology at a single cell level. The proposed platform provides a novel approach to high throughput and drug screening for different types of heart failure.
Different types of cardiac hypertrophy have been associated with an increased volume of cardiac myocytes (CMs), along with changes in CM morphology. While the effects of cell volume on gene expression are well known, the effects of cell shape are not well understood. This paper describes a method that has been designed to systematically analyze the effects of CM morphology on gene expression. It details the development of a novel single-cell trapping strategy that is then followed by single-cell mRNA sequencing. A micropatterned chip has also been designed, which contains 3000 rectangular-shaped fibronectin micropatterns. This makes it possible to grow CMs in distinct length:width aspect ratios (AR), corresponding to different types of heart failure (HF). The paper also describes a protocol that has been designed to pick up single cells from their pattern, using a semi-automated micro-pipetting cell picker, and individually inject them into a separate lysis buffer. This has made it possible to profile the transcriptomes of single CMs with defined geometrical morphotypes and characterize them according to a range of normal or pathological conditions: hypertrophic cardiomyopathy (HCM) or afterload/concentric versus dilated cardiomyopathy (DCM) or preload/eccentric. In summary, this paper presents methods for growing CMs with different shapes, which represent different pathologies, and sorting these adherent CMs based on their morphology at a single-cell level. The proposed platform provides a novel approach to high throughput and drug screening for different types of HF.
According to the World Health Organization, cardiovascular disease (CVD) is a major cause of morbidity and mortality worldwide. CVD dramatically affects the quality of people’s lives and has a huge socioeconomic impact. Cardiomyopathies, such as HCM and DCM, are primary disorders of the heart muscle and major causes of HF have been associated with high morbidity and mortality. There are many causes of HF, including environmental effects, such as infections and exposure to toxins or certain drugs8. HF can also be caused by genetic predisposition, namely mutations9. It is believed that the changes in genetic composition that affect extracellular matrix (ECM) molecules, integrins or cytoskeletal proteins could be responsible for impaired mechanosensation and various types of cardiac disease10.
The main feature of HCM is unexplained hypertrophy of the left ventricle11, and sometimes of the right ventricle12, and this frequently presents with predominant involvement of the interventricular septum. HCM is also characterized by diastolic dysfunction and myocyte disarray and fibrosis13. In most cases, the contractile apparatus of the heart is affected by mutations in sarcomeric proteins, leading to increased contractility of the myocytes14. In contrast, DCM is characterized by dilatation of one, or both, ventricles and has a familial etiology in 30% to 50% of cases15. DCM affects a wide range of cellular functions, leading to impaired contraction of the myocytes, cell death and fibrotic repair16.
Genetics has shown that certain types of mutations force single CMs to adopt specific shape characteristics during HCM3, namely square-shaped cells with a length:width AR that is almost equal to 1:14 (AR1). The same is true for DCM, with elongated cells with an AR that is almost equal to 11:1 (AR11). In addition, HF can be caused by increased afterload (e.g., in hypertension). In these cases, hemodynamic demands force CMs to take on square shapes, according to the Laplace’s law, and the AR changes from 7:15 (AR7) to 1:16,7. HF can also be caused by an increase in preload (e.g., in conditions that lead to volume overload). When this happens, the biophysical constraints force CMs to elongate and the AR changes from 7:1 to 11:1.
Signaling activity at membranes depend on global cell geometry parameters, such as the cellular AR, size, the membrane surface area and the membrane curvature18. When neonatal rat CMs were plated on substrates that were patterned to constrain the cells in a specific length:width AR, they demonstrated the best contractile function when the ratios were similar to the cells in a healthy adult heart. In contrast, they performed poorly when the ratios were similar to those of myocytes in failing hearts19. In the early stages of hypertrophy, cells become wider, as reflected by an increase in the cross-sectional area. HF occurs in the later stages of hypertrophy and cells typically appear elongated. Therefore, it is not surprising that in vivo rat models of chronic hypertrophy have reported an increase in the left ventricular myocyte length of around 30%20, but adult CMs from transgenic mouse model that were acutely treated with hypertrophic stimuli in vitro demonstrated similar increases in cell width instead21.
Single-cell RNA sequencing, which allows precise analysis of the transcriptome of single cells, is currently revolutionizing the understanding of cell biology. This technology was the preferred method when it came to answering the question of how did individual cell shapes affect gene expression. We compared single cells with different shapes, in particular with ARs of 1:1, 7:1 or 11:1. This was done by seeding the neonatal rat ventricular CMs onto a specially designed chip filled with the fibronectin-coated micropatterns2 with defined ARs of 1:1, 7:1 or 11:1. The micropatterns were fabricated using photolithography technology. The micropatterns were coated by fibronectin, surrounded by cytophobic surface. Therefore, CMs will attach, spread and capture the defined AR of micropatterns by solely growing on the fibronectin substrate, while avoiding the cytophobic area. The micropatterns are not in a well-shaped format. Instead, the fibronectin level is exactly at the same height of the surrounding cytophobic area. This provided similar conditions to growing cells in a Petri dish, as there is no stress from the surrounding walls. In addition, the surface area of micropatterns with different ARs are equal.
There were two particularly important aspects of the experimental design, which led to the use of single-cell RNA sequencing instead of bulk RNA sequencing. First, only a few percentages of the micropatterns can be occupied by a single cell. Second, sometimes a single cell does not fully occupy the micropattern surface. Single cells that completely cover a micropattern surface must be picked for single-cell RNA analysis. Because only a subgroup of the plated cells on a chip satisfied both criteria, it was not feasible to simply trypsinize the whole chip and collect all the cells for bulk RNA sequencing. Qualified cells needed to be picked individually using a semi-automated cell picker.
It currently remains unknown whether CM shape, by itself, has an intra-functional impact on the myocardial syncytium. The main purpose of the methods proposed in this paper was to develop a novel platform to study whether cell shape per se had an impact on the transcriptome17. Although in vitro studies are different from in vivo studies, the purpose of this study was to investigate the effect of different cell shapes on gene expression, bearing in mind that comparing cells with different shapes in vivo is extremely demanding. These experiments were inspired by Kuo et al.19, who used a similar approach and reported that they observed changes in physiological parameters due to changes in cell shape.
This study used single-cell RNA sequencing, which is a novel and powerful technology that can detect the transcriptome of single cells. It was combined with an innovative approach to culturing single CMs, so that they took on different ARs that, otherwise, could only have been observed in vivo.
The study had some limitations. For example, neonatal CMs had to be used to generate different morphotypes, as it is exceptionally challenging to culture enough vital adult CMs for 72 hours in defined shapes. Furthermore, CMs were cultured for 72 hours ex vivo, which might have had an impact on the gene expression pattern. However, this culturing was necessary, so that the cells could form specific morphotypes. Moreover, only single cells that were mononucleated and fully covered the fibronectin micropattern were selected for sorting. Patterned cells on each chip must be sorted merely in one round of sorting. Finally, about 50 cells that is roughly one-third of the selected cells were successfully picked up from each chip. There are two reasons that restricted the number of successfully picked cells. First, some cells were too tightly attached to the fibronectin pattern and the pickup flow was not forcible enough to successfully pick them up. Second, due to the trypsin treatment, the attachment between some cells and fibronectin became too loose. Consequently, these cells were pushed away from their fibronectin micropatterns, when the microcapillary approached them, and they were not picked up. The authors do not claim that this setup is the same as an in vivo environment, but it proved to be a viable approach to answering the research question.
The proposed method is applicable on different cell types (e.g., for hiPS-CMs). However, the following factors should be optimized to study other cell-types. Suitable ECM adhesive molecules for attachment of the specific cell-type should be used for coating the micropatterns. The geometry of the micropatterns should be modified according to the study question and cell-type. The culturing period can be modified based on the study question. The detachment reagent and its incubation time should be optimized precisely for the study cell-type. For instance, Accutase can be used instead of TryplE for detachment of embryonic and neuronal stem cells. The opening time parameters of the valves should be scrutinized to pick cells successfully, but gently. In summary, we engineered a novel platform to study cell shape that can provide a valuable resource for researchers in the field. In this context, we designed an experimental approach that mimicked in vitro characteristic shapes imposed on CM in vivo by hemodynamic constraints to identify the interplay between cellular architecture and gene expression. We also report the development of a novel platform to study HF in vitro and the identification of cell shape as a powerful determinant of gene expression. This is a novel observation with far-reaching implications for biology and medicine.
The authors have nothing to disclose.
None.
2100 Bioanalyzer Instrument | Agilent Technologies | G2939BA | Automated electrophoresis analyzer |
Anti-Red Blood Cell MicroBeads | Miltenyi Biotec | 130-109-681 | |
Axio Observer microscope | Zeiss | Z1 | Inverted microscope |
Betaine solution (5 M) | Sigma-Aldrich, MERCK | B0300 | |
CellSorter | CELLSORTER | https://www.singlecellpicker.com/ | Cell picker |
CYTOOchamber | CYTOO | 30-010 | Custom-designed chip |
CYTOOchip | CYTOO | 10-950-00-18 | Chamber |
DMEM | Thermo Fisher Scientific | 31966-021 | high glucose, GlutaMAX Supplement |
dNTP mix (25 mM) | Thermo Fisher Scientific | R1122 | |
Donkey Anti-Mouse IgG Alexa Fluor 488 | Abcam PLC | ab150105 | |
DTT (100 mM) | Thermo Fisher Scientific | 18064071 | |
ERCC RNA Spike-In Mix | Thermo Fisher Scientific | 4456740 | |
Fetal Bovine Serum | Thermo Fisher Scientific | 10082-147 | |
Fibronectin | Sigma-Aldrich, MERCK | F4759 | |
gentleMACS C Tube | Miltenyi Biotec | 130-093-237 | Rotor-cap tube |
gentleMACS Octo Dissociator with Heaters | Miltenyi Biotec | 130-096-427 | Dissociator with heater |
Greiner CELLSTAR Petri dish | Sigma-Aldrich, MERCK | P6987 | |
HEPES (1 M) | Thermo Fisher Scientific | 15630-056 | |
Horse Serum | Sigma-Aldrich, MERCK | H0146 | |
LD column | Miltenyi Biotec | 130-042-901 | |
Medium 199 | Thermo Fisher Scientific | 31150-022 | |
NE-1000 syringe pump | New Era Pump Systems | NE-1000 | Syringe pump |
Neonatal Cardiomyocyte Isolation Cocktail, rat | Miltenyi Biotec | 130-105-420 | |
Neonatal Heart Dissociation Kit, mouse and rat | Miltenyi Biotec | 130-098-373 | |
Oligo-dT30VN oligonucleotides | IDT Technology | 5′–AAGCAGTGGTATCAACGCAGAGTACT30VN-3′ | |
RNAse inhibitor (40 U µL-1) | Clontech | 2313A | |
sarcomeric α-actinin | Sigma-Aldrich, MERCK | EA-53 | |
SP8 confocal microscope | Leica Microsystems | SP8 | Confocal microscope |
Superscript II first-strand buffer (5x) | Thermo Fisher Scientific | 18064071 | |
Superscript II reverse transcriptase (200 U µL-1) | Thermo Fisher Scientific | 18064071 | |
Triton X-100 | Sigma-Aldrich, MERCK | T9284 | |
TryplE Express enzyme, no phenol red | Thermo Fisher Scientific | 12604013 | |
TSO (100 µM) | QIAGEN | 5′-AAGCAGTGGTATCAACGCAGAGTACATrGrG+G-3′ | |
Vibrant Dye Cycle green | Thermo Fisher Scientific | V35004 |