Water as a reactant in the differential expression of proteins in cancer

How the abundances of proteins are shaped by tumor microenvironments, such as hypoxic conditions and higher water content compared to normal tissues, is an important question for cancer biochemistry. Compositional analysis of more than 250 datasets for differentially expressed proteins compiled from the literature reveals a higher stoichiometric hydration state in multiple cancer types compared to normal tissue; this trend is also evident in pan-cancer transcriptomic and proteomic datasets from The Cancer Genome Atlas and Human Protein Atlas. These findings support the notion of a basic physicochemical link between increased water content in tumors and the patterns of gene and protein expression in cancer. The generally increased hydration state is juxtaposed with a wide spectrum of carbon oxidation states of differentially expressed proteins, which may be associated with different gene ages, host tissue properties and metabolic features of specific cancer types.

number of water molecules in these reactions (Table S1) was used as input to a residual analysis to culture experiments, open and filled symbols represent non-cancer and cancer cells, respectively, and squares represent yeast cells. Open symbols for cancer datasets represent mouse or rat models; all others are from human subjects. Dashed lines indicate the 50% credible region for highest probability density for all datasets for each condition. (C) Comparison of the 50% credible regions for cell culture and cancer tissue. Abbreviation: CRC -colorectal cancer. Table 2. Mean differences for all differential expression datasets in each condition, followed by log 10 of p-value in parentheses. p-values less than 0.05 (log 10 < -1.3) are shown in bold.  for which differential expression data were compiled in this study (color-coded circles in Fig. 3).

155
Differential gene expression for all cancer types taken together corresponds to significantly more 156 reduced proteins (Table 2, column DZ C ), but this is not evident in the HPA proteomics datasets. In    Table S2 for definitions 217 of abbreviations) compared to normal tissue, indicating generally higher expression of older genes.

218
I used a different calculation, where DPS represents non-weighted differences between the means 219 for up-and down-expressed genes, and obtained negative values for the same cancer types using 220 the TCGA/GTEx data (see Fig. 3C for selected cancer types that are paired between TCGA and HPA shows that many proteomes exhibit younger ages of the corresponding genes (DPS > 0) (Fig. 3C).

225
Taken together, the negative differences for transcriptomes are more statistically significant (Table 1). shown to consist for the most part of genes originating from the unicellular-multicellular transition of generally older genes that also code for more reduced proteins (Fig. 3 A and C reflects the hypoxic conditions that occur in many tumor microenvironments. 277 It is also somewhat surprising that hypoxia in cell culture generally does not induce the 278 up-regulation of more reduced proteins ( proteins secreted in hypoxia to be relatively oxidized (Table 2).

285
In marked contrast to the diverse trends of oxidation state, most cancer types are characterized by 286 a higher stoichiometric hydration state of proteins at both the transcriptional and translational levels.

287
These results indicate that water is consumed as a reactant when the differential expression of proteins

305
[270] found that gene expression patterns in aneuploid yeast cells are similar to those in normal 306 yeast cells exposed to hypoosmotic (that is, more dilute, the opposite of hyperosmotic) conditions.  Table 2). The present results support the hypothesis that osmotically induced dehydration provides a 328 thermodynamic drive for the preferential expression of proteins with lower stoichiometric hydration 329 state.

330
An initially unexpected finding is that the hydration state of proteins is substantially lower in 331 3D culture, including spheroids and aggregates, compared to traditional 2D culture in monolayers  [19] was also expanded in this study, but a fish gill proteome and two transcriptomic datasets were 367 excluded, and high-salt and high-glucose datasets were analyzed separately.

368
Lists of significantly differentially expressed proteins were taken directly from the original   Antibody staining intensities were converted to a semi-quantitative scale (not detected: 0, low: 1, 388 medium: 3, high: 5). The expression level score for each protein was calculated by averaging the score 389 for available samples, including "not detected" but excluding unavailable (NA) observations, and, for 390 normal tissues, observations in all available cell types. Differences in expression score between normal 391 and cancer ≥ 2.5 or ≤ -2.5 were considered to be differentially expressed proteins.