Engineering

Engineering

1. Introduction

Our project aimed to engineer a robust experimental model to study the role of NRAS mutations, particularly NRAS^G12D, in cancer progression and therapy resistance. Conventional 2D culture systems fail to capture the complexity of the tumor microenvironment, and simple gene overexpression models often lack stable, long-term expression. To address these limitations, we combined plasmid vector engineering with chip-based 3D culture technology.

Guided by the Design–Build–Test–Learn (DBTL) cycle, we iteratively refined plasmid constructs, transfection strategies, and chip architecture. Through this process, we established a platform that ensures stable NRAS expression and reproducible functional analysis within a physiologically relevant 3D-like invasion environment.

2. Plasmid Vector Engineering

<Design>

Figure 1
Figure 1 (Plasmid map of NRAS^WT construct. Major features include EF1α promoter, Kozak, NRAS, Luc2, mCherry, and puromycin resistance cassette)

We designed plasmids to constitutively express either wild-type NRAS (NRAS^WT) or the oncogenic mutant NRAS^G12D. The EF1α promoter was selected due to its robust and stable activity across mammalian cells, reducing the risk of silencing compared to CMV. A Kozak sequence was included to optimize translational initiation.

To enable co-expression of reporters, we introduced self-cleaving 2A peptides (T2A/P2A) between coding sequences, allowing a polycistronic transcript to generate multiple proteins. Luc2 provided a quantitative luminescence readout of cell viability and NRAS expression, while mCherry offered a fluorescent marker for imaging and tracking transfected cells. Transcript stability was ensured with a BGH/SV40 polyA signal, and a puromycin resistance cassette facilitated selection of stable NIH3T3 lines.

For the NRAS coding region, we used the human NRAS sequence (RefSeq: NM_002524.5). Two variants were synthesized:

  • Wild-type NRAS (NRAS^WT): containing the canonical glycine codon (GGT) at position 12.
  • Mutant NRAS (NRAS^G12D): carrying a single base substitution (GGT → GAT) at codon 12, which results in a glycine (G) to aspartic acid (D) change. This mutation is a well-known oncogenic driver in multiple cancers.

Full NRAS^WT and NRAS^G12D sequences are shown below.

Highlighted regions:

  • Yellow = primer binding sites used for PCR verification
  • Red = codon 12 (GGT → GAT) mutation site
NRAS^WT
ATGACTGAGTACAAACTGGTGGTGGTTGGAGCAGGTGGTGTTGGGAAAAGCGCACTGACAATCCAGCTAATCCAGAACCACTTTGTAGATGAATATGATCCCACCATAGAGGATTCTTACAGAAAACAAGTGGTTATAGATGGTGAAACCTGTTTGTTGGACATACTGGATACAGCTGGACAAGAAGAGTACAGTGCCATGAGAGACCAATACATGAGGACAGGCGAAGGCTTCCTCTGTGTATTTGCCATCAATAATAGCAAGTCATTTGCGGATATTAACCTCTACAGGGAGCAGATTAAGCGAGTAAAAGACTCGGATGATGTACCTATGGTGCTAGTGGGAAACAAGTGTGATTTGCCAACAAGGACAGTTGATACAAAACAAGCCCACGAACTGGCCAAGAGTTACGGGATTCCATTCATTGAAACCTCAGCCAAGACCAGACAGGGTGTTGAAGATGCTTTTTACACACTGGTAAGAGAAATACGCCAGTACCGAATGAAAAAACTCAACAGCAGTGATGATGGGACTCAGGGTTGTATGGGATTGCCATGTGTGGTGATG
NRAS^G12D
ATGACAGAGTACAAACTGGTGGTGGTGGGGGCCGACGGGGTGGGGAAGAGCGCCCTGACCATTCAGCTGATCCAGAATCACTTTGTGGACGAGTACGACCCAACCATCGAGGACAGCTATAGAAAACAGGTCGTGATCGATGGTGAAACCTGTCTGCTGGACATCCTGGACACTGCAGGCCAGGAGGAGTACTCTGCCATGAGAGACCAGTACATGAGGACCGGCGAGGGATTTCTGTGCGTGTTTGCTATCAACAACTCTAAGAGTTTCGCCGATATCAATCTGTACAGAGAGCAGATCAAAAGGGTGAAGGACAGCGACGACGTGCCCATGGTGCTGGTGGGCAATAAGTGCGACCTGCCAACCCGGACAGTGGATACCAAGCAGGCCCACGAACTGGCCAAGAGTTACGGGATCCCTTTCATCGAGACCAGCGCCAAGACCAGACAGGGCGTGGAGGACGCCTTCTACACCCTGGTGAGGGAGATTCGGCAGTACCGCATGAAGAAGCTGAATAGCTCCGACGACGGAACACAGGGCTGCATGGGCCTGCCCTGTGTGGTGATG

By verifying these sequence constructs with targeted primers, we ensured that our plasmids faithfully model both the wild-type and oncogenic states of NRAS, providing a reliable foundation for downstream functional assays.

<Build>

Plasmids encoding NRAS^WT and NRAS^G12D with an mCherry reporter were synthesized and introduced into NIH3T3 cells using Lipofectamine-mediated transfection. Transfection conditions followed the manufacturer’s recommended protocol (Thermo Fisher Scientific, “Transfecting Plasmid DNA into NIH3T3 Cells Using Lipofectamine® LTX Reagent”).

Stable pools were subsequently generated through puromycin selection (5 µg/ml for 2 days). This approach enriched reporter-positive cells and minimized heterogeneity for downstream assays.

To functionally model NRAS-driven phenotypes, these engineered plasmids were designed to express either wild-type or mutant NRAS^G12D upon nuclear delivery and transcription. Expression of NRAS variants allowed us to investigate proliferation, drug resistance, and migratory behavior in NIH3T3 cells, providing the foundation for downstream experiments within our lab-on-a-chip platform.

Figure 2
Figure 2 (Overview of plasmid vector transfection into NIH3T3 cells for NRAS construct delivery. Engineered plasmid vectors were introduced into NIH3T3 cells via lipid-based transfection. Once delivered into the nucleus, the plasmids undergo transcription to generate NRAS mRNA, which is subsequently translated into protein.)

<Test>

  • Selection with puromycin: To confirm the functionality of the puromycin resistance cassette in our plasmid, NIH3T3 cells were subjected to puromycin (5 μg/mL) for 2 days, beginning 48 h post-transfection.
  • Reporter validation: Brightfield and mCherry fluorescence imaging confirmed that NRAS^WT and NRAS^G12D constructs were successfully expressed in NIH3T3 cells compared with control.
Figure 3
Figure 3 (Puromycin selection of transfected NIH3T3 cells. NIH3T3 cells were transfected with plasmid and subjected to puromycin selection starting 48 h post-transfection. Brightfield images show: (left column) No puromycin treatment, (middle column) Cells treated with puromycin (5 µg/ml) for 2 days, and (right column) Cells after puromycin washout, showing only surviving resistant cells. Puromycin selection effectively eliminated non-transfected cells, leaving behind a population of resistant cells. Scale bar = 100 µm)
Figure 4
Figure 4 (Validation of plasmid expression in NIH3T3. Brightfield (top) and mCherry fluorescence (bottom) images of control (CON), NRAS^WT, and NRAS^G12D lines. Reporter fluorescence confirmed successful delivery and expression of engineered plasmids. Scale bar = 100 μm)

<Learn & Improve>

  • 1. Delivery optimization: Transfection efficiency was optimized by adjusting DNA-to-reagent ratios, ensuring high viability while maintaining robust reporter expression.
  • 2. Stable pools: Short and decisive puromycin selection (5 μg/mL for 2 days, followed by washout) generated uniform, reporter-positive populations, minimizing heterogeneity in assays.
  • 3. Lesson: Iterative adjustment of transfection and selection enabled us to build a robust, reproducible plasmid system where EF1α-driven expression produced stable lines, reliable reporters, and phenotype-specific outcomes. These plasmid-engineered cells formed the basis for subsequent chip assays.
  • 4. Improve: Beyond mCherry visualization, our plasmid system also contains a luciferase reporter, which can be leveraged in future assays for quantitative measurement of NRAS-driven phenotypes and drug responses. This dual-reporter strategy enhances flexibility—fluorescence allows rapid microscopy-based screening, while luminescence provides sensitive, quantitative readouts for downstream optimization.

3. Chip Engineering

<Design>

To better model tumor cell migration in a physiologically relevant context, we designed a microfluidic lab-on-a-chip (LOC) system with two compartments. Cells were seeded in the upper chamber, which was covered with a Matrigel layer to simulate the extracellular matrix barrier. The lower chamber collected migrated cells after they traversed through the Matrigel.

The connecting channels were filled with culture medium and placed on a rocking shaker, creating gentle oscillatory flow that mimicked body fluid dynamics. This setup allowed us to measure NRAS-dependent invasion under both ECM and dynamic fluid conditions.

Figure 5
Figure 5 (Schematic design of the lab-on-a-chip (LOC) device for migration assays. The chip consists of four layers: (1) a reservoir layer (8 mm holes, 4 mm thickness) for medium loading, (2) a top layer with 6 mm middle holes, 1 mm side holes, and 0.5 mm microchannels, (3) a bottom layer with 8 mm middle holes, 2 mm side holes, and 0.5 mm microchannels, and (4) a glass slide base for imaging)

<Build>

Figure 6
Figure 6 (Scheme of Lab-on-a-chip migration assay using Matrigel-embedded NIH3T3 cells.)

Chips were fabricated from PDMS via soft lithography, bonded to glass slides, and sterilized. Before seeding, the upper chamber was layered with Matrigel to form a 3D ECM barrier. NRAS^WT and NRAS^G12D NIH3T3 cells were seeded on top of this layer. The chip was supplied with medium in the lower chamber, and the entire system was placed on a rocking shaker, introducing gentle movement of the medium in the channels to simulate body fluid motion.

<Test>

  • Imaging: Brightfield microscopy confirmed migrated cells collected in the lower chamber after 72 h for control, NRAS^WT, and NRAS^G12D groups. Representative images showed markedly higher migration in NRAS^G12D compared with NRAS^WT and control.
Figure 7
Figure 7 (Brightfield images of migrated cells in the LOC assay after 72 h (CON, NRAS^WT, NRAS^G12D). Both NRAS^WT and NRAS^G12D showed higher migration than control, with NRAS^G12D exhibiting the strongest infiltration.)
  • Functional assays: Migrated cells were counted across multiple fields. Normalized to control (CON = 100%), NRAS^WT showed ~150–250%, and NRAS^G12D ~250–450% migration.
  • Dynamic conditions: rocking shaking supported nutrient exchange and physiological relevance.
  • Quantification: Both NRAS^WT and NRAS^G12D cells showed significantly increased migration compared with control, with NRAS^G12D showing the highest infiltration capacity (p < 0.01).
Table 1
Table 1 (Cells (CON, NRAS^WT, and NRAS^G12D) were seeded at 1×10^5 cells on Matrigel-coated LOC chambers. After 72 h, migrated cells in the lower compartment were counted from multiple microscopic fields and expressed as mean cell number ± SD. Data were normalized to the control group (CON = 100%). Both NRAS^WT and NRAS^G12D groups showed significantly higher migration compared to control, with NRAS^G12D exhibiting the strongest infiltration capacity (p < 0.05, *p < 0.01 vs. CON))

<Learn & Improve>

During the chip-based migration experiments, several technical challenges emerged:

  1. 1. Matrigel barrier thickness
    • ○ When the Matrigel layer was too thick, very few cells were able to infiltrate.
    • ○ When it was too thin, the barrier collapsed and failed to mimic a realistic ECM.
    • ○ Iterative adjustment led us to an intermediate thickness that allowed differential migration between groups while maintaining barrier stability for 72 h.
  2. 2. Rocking shaker speed
    • ○ Excessive shaking disrupted the Matrigel surface, causing irregular migration patterns.
    • ○ Insufficient shaking reduced nutrient exchange, which decreased cell viability.
    • ○ By calibrating the orbital speed, we reproduced physiologically relevant body fluid dynamics while keeping the ECM barrier intact.
  3. 3. Imaging and quantification
    • ○ Early counts varied due to inconsistent imaging fields.
    • ○ Standardizing the imaging area and normalizing migrated cell numbers to the control (CON = 100%) improved reproducibility across replicates.

Through optimization of ECM thickness, fluid dynamics, and imaging methods, we established a reproducible LOC invasion platform. This system consistently revealed that NRAS^G12D cells have markedly stronger migration capacity than NRAS^WT, faithfully reflecting oncogenic phenotypes in a 3D-like, dynamic environment.

4. Integration of Plasmid & Chip Platforms

The major innovation of our project lies in combining engineered NRAS plasmid vectors with a Matrigel-based LOC migration model. Stable NRAS^WT and NRAS^G12D lines were seeded at 1×10^5 cells per chip and migration was quantified after 72 h under rocking shaking.

This integration allowed us to:

  • Directly compare wild-type vs mutant NRAS phenotypes within the same microenvironment.
  • Quantify migration capacity using normalized brightfield counts.
  • Leverage reporter signals (mCherry, luciferase) to validate plasmid expression and support parallel functional assays.
  • Enable drug response evaluation under physiologically relevant ECM + fluid conditions.

By merging plasmid engineering with chip-based invasion assays, we created a robust platform to study oncogenic NRAS-driven phenotypes in a reproducible and translatable format.

5. Conclusion

By iteratively applying DBTL cycles, we engineered a composite platform that integrates plasmid vector design with microfluidic chip technology. This approach enabled the establishment of a stable, reproducible, and physiologically relevant system for investigating NRAS-driven oncogenesis.

Our findings demonstrate that precise genetic circuit design, combined with optimized chip engineering, can yield a scalable model for cancer biology and drug evaluation. This work underscores the potential of synthetic biology to bridge molecular-level engineering with tissue-level disease modeling, providing a versatile framework for future therapeutic discovery and translational research.