Association of computed tomography scan–assessed body composition with immune and PI3K/AKT pathway proteins in distinct breast cancer tumor components

Ting‐Yuan David Cheng, Dongtao Ann Fu, Sara M. Falzarano, Runzhi Zhang, Susmita Datta, Weizhou Zhang, Angela R. Omilian, Livingstone Aduse‐Poku, Jiang Bian, Jarome Irianto, Jaya Ruth Asirvatham, Martha Campbell‐Thompson

This study explores how a patient’s physical makeup—specifically the balance of fat and muscle measured via CT scans—relates to the internal molecular signaling of breast cancer tumors. By using spatially resolved protein profiling, the researchers found that low muscle mass and high body fat are associated with distinct changes in immune and growth-signaling pathways within the tumor microenvironment, suggesting that body composition may be a more precise tool for personalizing cancer care than BMI.

TL;DR

Researchers used spatially resolved proteomics to show that CT-assessed body composition (fat vs. muscle) is significantly associated with immune and PI3K/AKT signaling pathways in breast tumors. The findings suggest that "low muscle" and "high adiposity" phenotypes correlate with specific protein markers that could influence tumor progression and treatment response, outperforming traditional metrics like BMI.

What Problem Does This Paper Address?

While we know that obesity and sarcopenia (muscle loss) are linked to poor survival in breast cancer patients, the biological "why" remains murky. Traditional metrics like Body Mass Index (BMI) are blunt instruments; they don't distinguish between fat and muscle or tell us how a patient's body composition actually alters the tumor's behavior.

The challenge is that tumors are complex ecosystems. A protein expressed in the tumor cells (epithelium) might have a completely different clinical meaning than the same protein expressed in the surrounding support tissue (stroma). This paper addresses the gap in understanding how systemic body composition influences these distinct tumor compartments at the molecular level.

Methods

The study analyzed 52 patients with newly diagnosed breast cancer. The researchers combined two high-tech approaches:

  1. CT-Based Body Composition: Instead of BMI, they used clinical CT scans at the third lumbar vertebra (L3) to precisely quantify Total Adipose Tissue (TAT) and Total Skeletal Muscle (TSM). Patients were categorized into four types: Adequate, High Adiposity, Low Muscle, and High Adiposity with Low Muscle.
  2. GeoMx Digital Spatial Profiling (DSP): This is a multiplexed protein assay that allows researchers to "mask" specific areas of a tissue sample. They separated the tumor epithelium (the cancer cells) from the leukocyte-enriched stroma (the immune-rich surrounding area) to measure 43 different proteins in each compartment separately.

The team then used linear mixed models to correlate these body types with protein expression, adjusting for factors like age, race, and tumor stage.

Datasets

  • Primary Dataset: A cohort of 52 invasive breast cancer patients treated at the University of Florida Health Shands Hospital (2011–2020) with available CT scans and tumor tissue.
  • Validation/Contextual Data: The researchers used data from The Cancer Genome Atlas (TCGA) to perform survival analyses on the specific protein markers identified in their primary cohort.

Key Insights and Ideas

  • Spatial Specificity Matters: The associations found were compartment-specific. For example, the "low muscle" phenotype was linked to changes in the PI3K/AKT pathway within the tumor cells, but changes in immune cell markers were primarily seen in the stroma.
  • The "Low Muscle" Signature: Patients with low muscle mass showed higher expression of INPP4B (a protein that regulates PI3K signaling) and CTLA4 (an immune checkpoint). This suggests that muscle loss might be linked to a more immunosuppressive tumor environment.
  • The "High Adiposity" Signature: High body fat was associated with increased markers for T-cells (CD3, CD8) and B-cells (CD20) in the stroma. While often associated with inflammation, in some contexts, a "hotter" immune environment can actually lead to better responses to certain therapies—a potential molecular link to the "obesity paradox" seen in some cancer trials.
  • Beyond BMI: The study found that CT-derived metrics were stronger predictors of tumor protein expression than BMI or waist circumference, highlighting the need for more sophisticated body composition analysis in oncology.

Key Results

  • Low Muscle Type: Associated with significantly higher INPP4B in the stroma (log2-fold change = 1.14, p = 0.0003) and increased CTLA4 in the tumor epithelium. It was also associated with decreased pan-AKT expression.
  • High Adiposity Type: Correlated with increased stromal expression of CD3, CD8, CD20, and CD45RO (all p < 0.05), indicating a higher density of various immune cells.
  • High Adiposity with Low Muscle: This "worst of both worlds" phenotype was associated with lower expression of phospho-tuberin and phospho-PRAS40, both of which are key regulators of the MTOR growth pathway.

Limitations and Future Work

The authors acknowledge several limitations:

  • Sample Size: With only 52 patients, this is a "hypothesis-generating" study. The small numbers, particularly in the "high adiposity with low muscle" group (n=3), limit statistical power.
  • Timing: CT scans were "opportunistic," meaning they weren't all taken at the exact same time relative to diagnosis or surgery.
  • Diversity: The cohort was primarily hormone receptor-positive (ER+) tumors; results might differ for triple-negative or HER2+ subtypes.

Future Directions: The researchers suggest larger, confirmatory studies to see if these body composition "signatures" can predict which patients will respond best to immunotherapy or PI3K inhibitors. They also aim to investigate how physical activity interventions might "flip" these molecular switches within the tumor.

Summary updated

Abstract

Abstract This hypothesis-generating study aims to examine the extent to which computed tomography-assessed body composition phenotypes are associated with immune and PI3K/AKT signaling pathways in breast tumors. A total of 52 patients with newly diagnosed breast cancer were classified into four body composition types: adequate (lowest two tertiles of total adipose tissue [TAT]) and highest two tertiles of total skeletal muscle [TSM] areas); high adiposity (highest tertile of TAT and highest two tertiles of TSM); low muscle (lowest tertile of TSM and lowest two tertiles of TAT); and high adiposity with low muscle (highest tertile of TAT and lowest tertile of TSM). Immune and PI3K/AKT pathway proteins were profiled in tumor epithelium and the leukocyte-enriched stromal microenvironment using GeoMx (NanoString). Linear mixed models were used to compare log2-transformed protein levels. Compared with the normal type, the low muscle type was associated with higher expression of INPP4B (log2-fold change = 1.14, p = 0.0003, false discovery rate = 0.028). Other significant associations included low muscle type with increased CTLA4 and decreased pan-AKT expression in tumor epithelium, and high adiposity with increased CD3, CD8, CD20, and CD45RO expression in stroma (P&lt;0.05; false discovery rate &gt;0.2). With confirmation, body composition can be associated with signaling pathways in distinct components of breast tumors, highlighting the potential utility of body composition in informing tumor biology and therapy efficacies.
BibTeX
@article{cheng2024association,
  author = {Ting‐Yuan David Cheng and Dongtao Ann Fu and Sara M. Falzarano and Runzhi Zhang and Susmita Datta and Weizhou Zhang and Angela R. Omilian and Livingstone Aduse‐Poku and Jiang Bian and Jarome Irianto and Jaya Ruth Asirvatham and Martha Campbell‐Thompson},
  title = {{Association of computed tomography scan–assessed body composition with immune and PI3K/AKT pathway proteins in distinct breast cancer tumor components}},
  journal = {International Journal of Molecular Sciences},
  year = {2024},
  doi = {10.3390/ijms252413428},
}
Figure