Early selection of Chinese Hamster Ovary (CHO) cells using simultaneous label-free autofluorescence multi-harmonic (slam) microscopy and Fluorescence Lifetime Imaging Microscopy (FLIM) for biopharmaceutical cell line development

Efficient selection of high performing Chinese Hamster Ovary (CHO) clones remains a major bottleneck in biopharmaceutical development, requiring weeks to months of subculture and functional assays. Here, we introduce a label free, multimodal imaging pipeline that combines Simultaneous Label free Autofluorescence Multi-harmonic (SLAM) microscopy with Fluorescence Lifetime Imaging Microscopy (FLIM), coupled with a linear kernel Support Vector Machine (SVM) classifier, to enable early phenotypic classification of monoclonal CHO lines. By capturing endogenous two and three photon autofluorescence (NAD(P)H, FAD), second and third harmonic generation, and lifetime contrasts in a single acquisition, more 200 quantitative features, spanning intensity, shape, texture, and colocalization metrics, at the single cell level were extracted and analyzed. On a set of four producer clones (A–D), our SLAM + FLIM + SVM workflow achieves > 95 % balanced accuracy and per class AUC > 0.98 immediately at passage 2, vastly outperforming brightfield‐only models (≤ 0.6). The approach generalizes to an eight cell line panel (S1–S8), sustaining high (> 0.90) accuracy across passages P0–P5 and yielding robust unimodal benchmarks that confirm the synergy of combined multiphoton and lifetime contrasts. In addition, binning clones into low/medium/high classes for harvest titer, per‐cell productivity, maximum viable cell concentration, and long‐term stability, demonstrates early prediction of each critical quality attribute (CQA) (balanced accuracies 0.76–0.93) and identifies the most predictive features via permutation importance analysis. Importantly, rank weighted lifetime–lifetime and multiphoton–Redox colocalization metrics dominate, linking subcellular metabolic microarchitecture to functional performance. This study establishes the combination of SLAM and FLIM imaging as a high content phenotyping platform for label free CHO clone selection, with the potential to drastically shorten cell line development timelines.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Bioengineering

Grantor

University of Illinois Urbana-Champaign

Year dc:date

2025

Author dc:creator

• Ho, Alexander

Contributors dc:contributor

• Boppart, Stephen A

• Biophotonics
• Microscopy
• SLAM
• Simultaneous Label-free Autofluorescence Multi-harmonic
• FLIM
• Fluorescence Lifetime Imaging Microscopy
• Multimodal
• Imaging
• Fluorescence
• Multiphoton
• Nonlinear Microscopy
• Redox
• CLD
• Cell line development
• Pharmaceutical
• Vaccine
• CHO
• Chinese Hamster Ovary
• live cell imaging

Statement dc:rights

• © 2025 Alexander Ho

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/129578

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F129578