How do pre-existing resistant subpopulations differ from metastatic tumors?
We extensively examined how certain subsets of cancer cells exhibit intrinsic resistance to therapy. These cells may harbor specific mutations in drug targets or downstream pathways, or exhibit transcriptional programs that enable them to survive, continue proliferating, and actively efflux anti-cancer drugs.
Metastatic cancer cells originate from tumor cells that escape the primary site, survive the stress-induced selection pressures of circulation, and colonize new tissues. While they share some molecular pathways with resistant, in situ subpopulations, metastatic cells often evolve further, enhancing their capacity to migrate, establish themselves in new niches, evade local immune surveillance, and resist treatment.
Therefore, even though both cell types originate from the same tumor, they respond differently to therapy. Consequently, the use of predictive biomarkers becomes critical for guiding treatment planning and patient stratification.For instance, in this study, the authors used scRNA-seq with combinatorial barcoding to analyze the mode of invasion in a patient-derived model of head and neck squamous cell carcinoma (HNSCC). They found distinctive signatures in the single cell migration and collective migration modes, discovering that the collective signature is a survival predictor and therefore could be used as a prognostic tool.
Another study investigated how loss of Glutathione-S-transferase Theta 2 (GSTT2) affects non-muscle invasive bladder cancer (NMIBC) response to immunotherapy. NMIBC has a high recurrence rate that may eventually progress in muscle-invasive disease, and a GSTT2 promoter deletion has been linked to better response to immunotherapy, preventing recurrence and muscle invasion. The authors found that GSTT2 loss influences tumor response to immunotherapy mainly by altering expression of genes tied to cancer progression and immune modulation. Given the deletion is common in humans, GSTT2 may be a useful biomarker for evaluating and predicting patient responses to BCG immunotherapy.
Understanding how the Tumor microenvironment contributes to resistance and how it changes in response to therapy.
Cancer cells are not the only players in the drug resistance game.
The tumor microenvironment is composed of stromal elements such as CAFs diverse immune populations including TAMs, T lymphocytes, and NK cells, as well as the ECM. These components actively contribute to tumor initiation, progression, and therapy resistance.
Early therapeutic strategies aimed at targeting the TME were based on the premise that stromal and immune cells, being genetically stable, would be less prone to developing drug resistance, but recent findings reveal that tumors in different organs display tissue-specific microenvironments that can alter treatment responses in unexpected ways.
Chemotherapy and other targeted therapies can indeed initially shift the TME toward a more immunostimulatory state but these changes are often short-lived. As treatment continues, adaptive responses in the TME can counterbalance these effects: CAF activation increases, ECM deposition intensifies, and immunosuppressive myeloid populations expand, ultimately reinstating tumor-protective conditions.
Sc-RNA seq is uniquely positioned to study the interplay within the TME to track how cells shift during cancer treatment.
In breast cancer, CAFs and macrophages are key stromal and immune components that drive tumor progression through matrix remodeling, immunomodulation, and promotion of proliferation and invasion. To study their synergistic effects, researchers developed a 3D breast tumor microenvironment-on-a-chip (TMEC) model with scRNA-seq in triple-negative breast cancer (TNBC). The analysis revealed that CAFs and macrophages jointly enhance tumor aggressiveness and identified KYNU-driven activation of the kynurenine pathway as a potential immune evasion mechanism, suggesting its pharmacological inhibition as a therapeutic avenue.
In pancreatic cancer recent advances in high-resolution single cell and spatial transcriptomics reveal how tumors resist therapy by both altering cellular states and rewiring the TME. In pancreatic ductal adenocarcinoma, a single cell transcriptome analysis of the heterogeneity of the classical subtype pancreatic adenocarcinoma (PDAC) epithelium, researchers identified four distinct cell clusters that appeared, in varying proportions, in all tumor samples analyzed. One of the clusters consistently displayed a basal-like gene expression profile, typically associated with a more aggressive tumor behavior.
Spatially resolved transcriptomics showed that neoadjuvant therapy reshaped CAF–tumor communication, reducing overall ligand–receptor pairs but enriching IL-6 family signaling. Functional assays confirmed IL-6’s role in chemoresistance, illustrating how treatment-induced TME remodeling fosters drug-tolerant niches. Together, these findings highlight two complementary resistance mechanisms—cell-intrinsic adaptations like endocycling and TME-driven signaling rewiring—and demonstrate how combining combinatorial barcoding scRNA-seq with spatial interaction modeling can expose new therapeutic targets such as IL-6 blockade or endocycling pathway inhibition.Another study using scRNA-seq with combinatorial barcoding profiled a glioblastoma cancer cell line after chemotherapy and discovered a rare endocycling cell state (cells that replicate DNA without division).
These cells were larger, had dramatically increased RNA content, and followed distinct transcriptional trajectories from untreated cells, suggesting a rewired, drug-resistant state that evades mitotic checkpoints. This phenotype persisted under treatment, indicating a potential reservoir for relapse. Parse’s combinatorial barcoding uniquely captured these atypical, large cells, which are often missed by droplet-based scRNA-seq, highlighting its value for detecting rare resistant populations that may also occur in the TME.
Ask a Single Cell
Therapeutic resistance is a major obstacle to durable cancer treatment. While many therapies reduce tumor burden, a subset of cancer cells can evade treatment through intrinsic or adaptive mechanisms, ultimately driving relapse and progression. These resistant populations are often rare, dynamic, or transcriptionally plastic, making them difficult to capture. Single cell RNA sequencing enables the detection and characterization of resistant cell states, distinguishing pre-existing and therapy-induced programs and clarifying how cancer cells and the tumor microenvironment evolve under therapeutic pressure.
Specific cellular pathways activated in response to therapies
- Do all tumor cells activate the intended pathway, or only a subset?
- Are cancer stem-like cells transcriptionally resistant to the drug?
- Does drug-induced pathway inhibition differ across tumor subclones?
- Are there pre-existing expression states that predict drug-induced pathway activation?
How different cancer cell states respond to therapeutics, and what mechanisms could underpin these responses
- Are aggressive or therapy-relevant states (stem-like, EMT) rare or dominant?
- Which cell states are selectively eliminated by treatment?
- Which states persist or expand under therapeutic pressure?
- Are resistant states pre-existing or newly induced?
What are the off-target effects of novel chemotherapeutics
- Which non-tumor cell types show transcriptional responses to treatment?
- Are immune or stromal cells directly affected by the drug?
- Are stress or damage pathways activated in non-target cells?
- Are unintended signaling pathways consistently activated?
- Which off-target responses appear shortly after treatment?
- Are early transcriptional changes predictive of later toxicity?
How do pre-existing resistant subpopulations differ from metastatic tumors
- Are metastatic transcriptional programs already present in the primary tumor?
- Are resistant cells a subset of metastasis-like states, or a separate class?
- Do resistant cells rely more on stress tolerance than growth?
- Are metastatic cells more proliferative or metabolically active?
- Do resistant cells rely on protective local niches?
- Can early elimination of resistant states prevent metastasis?
Understanding how the tumor microenvironment contributes to resistance and how it changes in response to therapy
- Does treatment create new resistance-supporting niches?
- Are therapy-induced TME changes transient or stable?
- Do resistant tumor cells induce suppressive immune states?
- Are stromal cells transcriptionally reprogrammed by resistant tumors?
- Are resistant tumors associated with specific suppressive immune states?
- How does immune composition change in non-responders?
TLDR: Cancer cell populations are inherently primed to evolve under pressure. When therapy is applied, most cells are eliminated, but a subset can adapt, persist, or shift state in ways that enable survival. These resistant populations may preexist or arise through therapy-induced transcriptional and microenvironmental changes, ultimately driving relapse and disease progression. Single cell RNA sequencing makes these evolutionary responses visible, revealing how resistant states emerge, expand, and interact with the tumor microenvironment during treatment. This chapter explores how understanding resistant subpopulations informs biomarker discovery and strategies to prevent therapeutic failure.
