🧬 biology

PGK1 expression differs across breast cancer molecular subtypes and is associated with overall survival in the TCGA BRCA cohort

This study demonstrates that PGK1 expression varies significantly across breast cancer molecular subtypes and serves as an independent predictor of shorter overall survival in the TCGA-BRCA cohort after adjusting for age, stage, and subtype.

Original authors: He Bu, Junwen Bai

Published 2026-08-03

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: He Bu, Junwen Bai

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

Imagine your body as a bustling city where every cell is a worker. Usually, these workers follow a strict, efficient power grid to get their energy, much like a city running on a reliable, steady electricity supply. But sometimes, when a neighborhood gets chaotic—like in cancer—the workers switch to a frantic, messy backup generator. This backup system is called "glycolysis." It's a fast but inefficient way to burn sugar for energy, and it's a favorite trick of many cancer cells to grow quickly and ignore the rules. Scientists have long suspected that one specific part of this messy backup generator, a tool called PGK1, might be a key player in how aggressive breast cancer gets. But here's the tricky part: breast cancer isn't just one disease. It's more like a city with four very different neighborhoods (called molecular subtypes), each with its own unique culture, speed, and rules. If you look at the city's chaos without realizing which neighborhood you're in, you might get the wrong idea about what's causing the trouble. So, the big question is: Is PGK1 actually a dangerous troublemaker on its own, or is it just loud because it lives in a particularly chaotic neighborhood?

This study decided to play detective with a massive digital map of 1,082 breast cancer tumors from a public database called TCGA. The researchers wanted to see if the amount of PGK1 in a tumor could predict how long a patient would survive, but they were careful to check if the "neighborhood" (the cancer subtype) was messing up the clues. They found that PGK1 expression is indeed very different across these neighborhoods. It's like a volume knob that is turned way up in the "HER2-enriched" and "Basal-like" neighborhoods, while it's turned down in the "Luminal A" and "Normal-like" areas.

When they looked at the patients, the story got a bit more serious. The data showed that patients with higher levels of PGK1 tended to have shorter overall survival times. Think of it this way: if PGK1 were a speedometer, a higher reading was linked to a faster, more dangerous ride. Even after the researchers carefully adjusted for the patient's age, the stage of the cancer, and which specific neighborhood the tumor lived in, the link remained. In their main analysis, for every standard step up in PGK1 levels, the risk of death increased by about 41%. This suggests that PGK1 isn't just a bystander; it might be a genuine marker of danger, even when you account for the type of cancer.

The researchers also peeked at what other genes were hanging out with PGK1. They found it was best friends with other genes involved in that messy sugar-burning process (glycolysis) and genes related to rapid cell division. It was like finding PGK1 at a party with the most energetic, chaotic guests. Interestingly, it was less likely to be seen with genes that usually keep cells calm and organized.

However, the authors are very careful not to shout "We found the cure!" or "This is the cause!" They emphasize that this is a single study looking at past data, so it's more of a strong clue than a final verdict. They didn't test this in a lab or on new patients, and they didn't prove that PGK1 causes the cancer to be deadly, only that the two are linked. They also noted that they couldn't see exactly which cells in the tumor were making the PGK1, just the average from the whole lump. So, while the findings are exciting and suggest PGK1 is a promising target for future investigation, the paper concludes that we need more independent studies and experiments to confirm if this tool is truly the villain or just a very loud witness in the story of breast cancer.

Technical Summary: PGK1 Expression and Survival in Breast Cancer Subtypes

Problem Statement
Phosphoglycerate kinase 1 (PGK1) is a glycolytic enzyme implicated in cancer metabolism and breast cancer progression. While prior studies have suggested associations between elevated PGK1 expression and unfavorable clinical outcomes, these findings have been inconsistent. The authors posit that previous reports may have been confounded by the failure to account for the distinct molecular subtypes of breast cancer (e.g., PAM50 classifications) and disease stage. Furthermore, methodological limitations in prior work, such as the use of dichotomized expression data, incomplete clinical adjustments, and heterogeneous data pipelines, have cast uncertainty on the robustness of the PGK1–survival association.

Methodology
This study conducted a retrospective, single-cohort analysis using the official TCGA-BRCA PanCancer Atlas 2018 dataset obtained from the cBioPortal DataHub. The analysis focused on 1,082 primary tumor samples with RNA-seq V2 RSEM expression data.

Data Preprocessing: Expression values were log2-transformed. Clinical data included age, AJCC pathologic stage, and overall survival (OS) status. PAM50 subtypes were mapped to Luminal A, Luminal B, HER2-enriched, Basal-like, and Normal-like categories.
Survival Analysis: A complete-case cohort of 915 patients (with 118 deaths) was utilized for survival modeling. PGK1 expression was treated as a standardized continuous variable. Three Cox proportional hazards models were fitted:
1. Univariate (PGK1 only).
2. Age-adjusted.
3. Primary Model: Age-adjusted with the baseline hazard stratified by AJCC stage and PAM50 subtype. This stratification was chosen to address non-proportional hazards observed in categorical models for stage and subtype.
Co-expression Analysis: A curated 119-gene panel was used to calculate composite signatures for glycolysis, HIF-1 regulation, epithelial–mesenchymal transition (EMT), and cell cycle. Pearson and partial correlations (adjusted for PAM50) were computed between PGK1 and these signatures.
Statistical Rigor: Analyses were performed using Python. The study adhered to REMARK guidelines, utilized two-sided tests, and provided full reproducibility scripts and checksums.

Key Results

Subtype-Specific Expression: PGK1 expression varied significantly across PAM50 subtypes ( $P = 4.35 \times 10^{-57}$ ). Expression was highest in HER2-enriched tumors (mean 13.76), followed by Basal-like (13.41), Luminal B (13.16), Normal-like (12.82), and Luminal A (12.76).
Survival Association: Higher PGK1 expression was significantly associated with shorter overall survival.
- In the univariate model, the hazard ratio (HR) per 1-SD increase was 1.55 ( $P = 7.10 \times 10^{-7}$ ).
- In the primary stratified model (adjusting for age, stage, and subtype), the association remained statistically significant with an HR of 1.41 (95% CI, 1.13–1.76; $P = 0.002$ ). This indicates a 41% increase in hazard per standard deviation of PGK1 expression, independent of stage and molecular subtype.
Co-expression Patterns: PGK1 showed the strongest positive correlation with a glycolysis composite score ( $r = 0.675$ ; partial $r = 0.575$ after subtype adjustment). It also correlated positively with cell-cycle and HIF-1 regulatory scores. Conversely, PGK1 was negatively correlated with luminal hormone-receptor markers (ESR1, PGR) and BCL2.

Significance and Claims
The authors conclude that PGK1 expression is not uniform across breast cancer but varies substantially by molecular subtype, with the highest levels found in aggressive HER2-enriched and Basal-like tumors. Crucially, the study demonstrates that the association between high PGK1 expression and poor overall survival persists even after rigorous adjustment for age, AJCC stage, and PAM50 subtype.

The paper positions these findings as exploratory. It explicitly states that while the data supports PGK1 as a potential biomarker for risk stratification, the study does not establish causal regulation, clinical utility, or therapeutic relevance. The authors emphasize that these single-cohort findings require independent clinical validation and functional experimental verification before any clinical conclusions can be drawn. The study's primary contribution lies in its methodological rigor—using official source files, continuous expression modeling, and explicit handling of non-proportional hazards—to clarify the independent prognostic value of PGK1 in the context of breast cancer heterogeneity.

Technical Summary: PGK1 Expression and Survival in Breast Cancer Subtypes

More like this