Results- - Proof of Concept

Experiment 1: HDAC6 knockdown validation

To verify the functionality of our therapeutic RNA interference design, we conducted an initial validation of HDAC6 silencing in K562 cells.

Objective: To confirm the efficiency and specificity of our miR-E-based therapeutic construct in downregulating HDAC6 expression.

Methods overview: Lentiviral particles carrying the EF-1α-driven miR-E shRNA targeting HDAC6, along with an EGFP reporter and puromycin resistance cassette, were produced and used to transduce K562 cells. Transduction efficiency was assessed through EGFP fluorescence.

Results: Transduction was successful, as evidenced by strong EGFP fluorescence in treated cells (Fig. 1A).

RT-qPCR analysis revealed a 4.02-fold (≈75.1%) reduction in two biological replicates (Fig.1B) in HDAC6 mRNA levels compared to untransduced controls, confirming robust transcriptional knockdown.

Flow cytometry analysis across two independent biological replicates further demonstrated a 30-35% decrease in HDAC6 protein levels, validating suppression at the protein level (Fig.1E, 1F).

To verify specificity, α-tubulin was used as a housekeeping marker. Tubulin expression remained unchanged in both replicates, indicating that the knockdown selectively targeted HDAC6 without affecting unrelated proteins (Fig.1C, 1D).

Since additional stock of transduced cells was available, we proceeded with immunofluorescence staining to visualize HDAC6 expression. The analysis revealed a clear reduction in HDAC6 signal intensity in cells treated with our construct compared to untransduced cells, further confirming effective silencing (Fig.1G).

Conclusion: The miR-E shRNA construct achieved efficient and specific knockdown of HDAC6 at both the mRNA and protein levels, as confirmed by qPCR, flow cytometry and immunofluorescence. This experiment lays the groundwork for the subsequent validation of our therapeutic concept.

Figures

Figure 1A. EGFP fluorescence confirming successful transduction.

Figure 1B. RT-qPCR showing HDAC6 mRNA reduction in K562 cells transduced with miR-E shRNA at 100 μL and 200 μL virus/mL.

Figure 1C. Flow cytometry plots showing α-tubulin (AF647) expression. Results represent two independent experiments.

Figure 1D. Quantification from two independent experiments showing the percentage of α-tubulin-positive cells. Knockdown of HDAC6 did not alter the total α-tubulin protein levels.

Figure 1E. Flow cytometry plots showing HDAC6 (PE) expression. Results represent two independent experiments.

Figure 1F. Quantification from two independent experiments showing the percentage of HDAC6 - positive cells. Results show reduction of HDAC6 protein upon knockdown (KD) in K562 cells.

Figure 1G. K562 cells transduced with 200 µL lentiviral miR-E-based shRNA targeting HDAC6 show reduced HDAC6 immunofluorescence compared to untransduced (UT) cells.

Experiment 2: AND Gate & Fail-safe system validation

Objective: To validate the dual-input AND gate logic and the integrated fail-safe mechanism containing four tandem miR-7-5p target sites, demonstrating selective circuit activation and miRNA-dependent repression under physiologically relevant conditions.

Methods overview: To simulate the layered logic of Morphe’s therapeutic system, we built a dual-inducible visual circuit composed of two complementary plasmids:

• Plasmid A: Drives the reporters TagBFP2 and mCherry under doxycycline- and cumate-inducible promoters, respectively.
  
• Plasmid B: Provides the constitutive expression of the rtTA and CymR regulators required for the two induction systems.
  

Together, these plasmids operate as a biological logic processor, expressing blue (TagBFP2), red (mCherry) or both (purple) depending on the combination of inducers.

Because lentiviral packaging was time-limited, both plasmids were transfected in HEK293T cells with HBS (HEPES-buffered saline) and CaCl₂. Each plasmid expressed a distinct antibiotic resistance gene (blasticidin and puromycin), allowing dual selection (10 µg/mL blasticidin + 2 µg/mL puromycin) to enrich double-positive cells.

Induction Matrix

Notes: No combinations of mimic and siRNA or triple-inducer setups were included in this experiment. Doxycycline was intentionally omitted from all mimic and siRNA conditions, as the mimic is antibiotic sensitive. Each condition was tested in three technical replicates to ensure reproducibility and reliable comparison across treatments.

Results: Confocal microscopy revealed distinct and reproducible fluorescence patterns corresponding to inducer combinations.

Control cells showed little to no detectable fluorescence, confirming minimal background activity. TagBFP2 was visible upon no induction indicating low basal leaky expression (Fig.2A, upper left).

Single induction with doxycycline or cumate activated TagBFP2 (blue) or mCherry (red), respectively (Fig.2A second row).

Dual induction activated both reporters: the purple signal observed represents the merged overlap of blue and red fluorescence channels in the confocal composite, confirming correct AND-gate behavior (Fig.2A, down right 3rd row).

In miR-7-5p mimic conditions, mCherry fluorescence driven by the cumate promoter was completely absent, while TagBFP2 expression remained unaffected (Fig.2B, 3rd and 4th row, 2nd column).

In contrast, scramble siRNA-treated cells retained detectable red fluorescence (Fig.2A, 2nd row, 2nd column), though weaker than in the cumate-only condition. This attenuation was attributed to cellular stress rather than incomplete repression, since the cells underwent CaCl₂ transfection, dual antibiotic selection, and serum-free exposure in Opti-MEM during transfection. To mitigate this, the mimic incubation period was extended to 48 h (followed by 24 h induction) to allow recovery after medium replacement.

Overall, the circuit was activated only under dual induction and effectively repressed in the presence of miR-7-5p, validating both the precision of the AND gate and the reliability of the fail-safe mechanism.

Conclusion: The dual-inducible circuit successfully demonstrated AND gate behavior, being activated only under simultaneous doxycycline and cumate induction. The integrated fail-safe module responded precisely to miR-7-5p mimic treatment, completely suppressing mCherry expression while leaving TagBFP2 unaffected. These findings confirm the expected logic and specificity of the designed regulatory system, although extended quantitative studies will be required to evaluate long-term stability and reliability.

TagBFP2 fluorescence was detectable without doxycycline induction (Fig. 2A) , indicating low basal expression under control conditions. This likely reflects the TRE-YB_TATA promoter architecture, cellular stress responses and the use of regular FBS instead of Tet-free FBS, rather than genuine promoter leakiness.

Limitations: This experiment was performed under short-term transient transfection, preventing long term evaluation. Follow-up assays are required to assess signal stability, reversibility, and durability under sustained induction and eventual lentiviral integration.

Figures

Figure 2A. HEK293T cells transfected with the dual-plasmid system (TRE_YB-TATA_TagBFP2 / rtTA_CymR).

Fluorescence images show uninduced cells, Dox induction, and combined Dox + cumate induction. Dox activates TagBFP2, cumate induces mCherry, and co-induction leads to magenta/purple co-expression.

Figure 2B. HEK293T cells transfected with the miRNA mimic upon cumate induction, compared with untreated cells and those transfected by the scramble siRNA.

Experiment 3: HDAC6 KD and insulin pathway response

Objective: To evaluate whether HDAC6 knockdown could promote a metabolic rescue of insulin signaling under obesity-mimicking, lipotoxic conditions induced by palmitic acid (PA).

Method overview: K562 cells were treated with PA at 100 µM (data not shown) and 200 µM , followed by serum starvation and insulin stimulation. Following treatment cells were fixed and stained for phospho-FoxO1 (p-FoxO1) to evaluate insulin-pathway activation by immunofluorescence.

Each concentration was combined with three viral transduction levels (UT, 100 µL virus, 200 µL virus), resulting in six total conditions:

• UT + 100 µM PA + starvation + insulin
• 100 µL virus + 100 µM PA + starvation + insulin
• 200 µL virus + 100 µM PA + starvation + insulin
• UT + 200 µM PA + starvation + insulin
• 100 µL virus + 200 µM PA + starvation + insulin
• 200 µL virus + 200 µM PA + starvation + insulin

Each condition was performed in two technical replicates.

Results: p-FoxO1 staining was successfully detected in all groups, confirming antibody performance and reproducibility between replicates.

However, no measurable change in p-FoxO1 signal intensity was observed between HDAC6-knockdown and untransduced (UT) cells at either palmitate concentration.

Similarly, no localization shift of p-FoxO1 was detected, indicating that HDAC6 silencing did not modify insulin-pathway activation in this cellular context.

Confocal imaging was performed only for the 200 µL virus groups, which showed the highest transduction efficiency the 100 µL virus conditions were not imaged but appeared similar under fluorescence microscopy.

Conclusion: This assay showed no detectable change in p-FoxO1 levels or localization following HDAC6 knockdown, suggesting that insulin-pathway activation was unaffected in K562 cells under lipotoxic stress.

Nevertheless, the experiment confirmed the technical feasibility of integrating viral transduction, palmitate treatment, and insulin stimulation in suspension cultures. K562 cells are leukemic and predominantly glycolytic (Warburg effect), meaning they exhibit minimal insulin receptor-mediated signalling, which likely masks potential rescue effects.

Future experiments will focus on insulin-responsive models (e.g. 3T3-L1 adipocytes, HepG2 hepatocytes) using quantitative readouts such as p-Akt (Ser473) to reveal any true metabolic rescue.

Limitations: The use of K562 cells, which exhibit the Warburg effect and limited insulin receptor activity, restricts the ability to model canonical insulin-responsive pathways.

Therefore, these findings primarily reflect technical feasibility rather than physiological recovery of insulin sensitivity.

Follow-up studies should incorporate quantitative assays (e.g. p-Akt or glucose uptake measurements) in metabolically active cell lines to confirm HDAC6’s role in restoring insulin responsiveness.

Figures

Figure 3A. K562 cells pre-treated with palmitic acid (200uM), serum-starved, acutely stimulated with insulin (100nM). Cells are stained with primary p-Foxo1 and secondary AF 647 antibodies.