Mouse Studies Find No Toxicity or Allergenicity Signals for Recombinant Soybean Hemoglobin

Plant-based burgers have long struggled with a stubborn problem: even when the texture comes close to beef, the color and the bloody, meaty flavor often do not. Hemoglobin, the heme-rich protein that gives animal meat its red hue and characteristic aroma, is the key ingredient that many alternative-meat makers have borrowed to close that sensory gap. A study published in the journal Food, Nutrition and Health now reports the results of a laboratory safety evaluation of recombinant soybean hemoglobin, a version of the protein produced with a food-grade yeast expression system, and finds no apparent acute or subacute toxicity and no significant sensitization potential in the tests performed.

The research, conducted by Nan Zhang, Jiayi Liu and colleagues at Huazhong Agricultural University, compared recombinant soybean hemoglobin with bovine hemoglobin, a reference protein drawn from animal blood, in a series of standardized mouse experiments. The work was motivated by a practical bottleneck in the alternative-protein industry. Hemoglobin for plant-based meat has traditionally been extracted from animal blood, which undermines the environmental and ethical rationale of the products, while microbial production routes have raised their own questions about whether the resulting protein is safe to eat in meaningful quantities over long periods.

The team’s production strategy builds on recent advances in microbial heme biosynthesis. Previous work by other groups achieved elevated heme levels in Escherichia coli, and food-grade microorganisms such as Saccharomyces cerevisiae, ordinary baker’s yeast, have been engineered to express functional hemoglobin through enhanced heme biosynthesis pathways. One earlier commercial route used the yeast Pichia pastoris, but that organism is not considered food-grade, so its safety requires additional scrutiny. The new study used recombinant soybean hemoglobin produced in a food-grade Saccharomyces cerevisiae system, prepared as a suspension in distilled water and delivered to the animals by oral gavage.

The acute phase of the testing involved forty-eight SPF-grade BALB/c mice, divided equally by sex into four groups: a control group receiving only water, and three groups receiving a single oral dose of recombinant soybean hemoglobin at 125, 250 or 500 milligrams per kilogram of body weight. Over the following fourteen days, researchers observed the animals daily for appearance, respiration, behavioral activity, food and water intake, oral and nasal secretions, urine and feces, and recorded mortality and body weight. No abnormalities were observed in any of these measures, no deaths occurred, and body weights did not differ significantly from the control group. Dissection of the major organs, including the liver, spleen and heart, revealed no overt abnormalities.

The subacute phase was longer and more detailed. Sixty mice, again balanced by sex, received daily oral gavage for twenty-eight consecutive days: one group received bovine hemoglobin at 400 milligrams per kilogram, one received recombinant soybean hemoglobin at the same dose, and a control group received water. No mortality occurred in any group, no signs of toxicity were noted, and body weight gain, food intake and water consumption were statistically indistinguishable from the control group. At the end of the trial the researchers weighed nine organs, including the heart, liver, spleen, lungs, kidneys, stomach, brain, testes and ovaries, and calculated organ coefficients relative to body mass.

Two statistically significant differences did emerge in the organ data, and the authors were careful to interpret them cautiously. The spleen coefficient in male mice of the bovine hemoglobin group and the brain coefficient in male mice of the recombinant soybean hemoglobin group were both significantly lower than in controls, but the values remained within the normal physiological range, and no other organ coefficients differed significantly. In female mice, none of the hemoglobin treatments produced significant changes in any organ coefficient. The researchers concluded that these isolated variations lacked toxicological relevance, a judgment supported by the complete blood count, which showed no significant differences in any hematological parameter across the treated groups after twenty-eight days.

Serum biochemistry told a similar story. Liver function indicators, including alanine aminotransaminase, aspartate aminotransaminase, alkaline phosphatase and albumin, showed no significant differences in any hemoglobin group. Metabolic indicators, covering triglycerides, total cholesterol, high-density lipoprotein cholesterol and glucose, were likewise unchanged. The one statistically significant finding was an elevated creatinine level in female mice of the bovine hemoglobin group, which the authors note remained within the normal range; renal indicators in the other groups did not differ from controls. The cardiac marker creatine kinase-MB showed no significant changes, and histopathological examination of hematoxylin and eosin stained sections of the heart, liver, spleen, lungs, kidneys, stomach, brain, small intestine and reproductive organs found no pathological alterations such as inflammatory infiltration, cellular necrosis or fibrosis in any treated group.

The study also went beyond conventional toxicology by examining the gut microbiome, which the authors describe as a sensitive biomarker of host health capable of reflecting subclinical physiological responses to exogenous substances. Fresh fecal samples collected at the end of the twenty-eight-day trial were sequenced and analyzed for alpha diversity, using indices including Sobs, Ace, Chao, Shannon and Simpson, and for beta diversity through principal component analysis. The results showed no statistically significant differences between the recombinant soybean hemoglobin group and controls in any diversity index, no clear separation of groups in the PCA plot, and highly consistent community composition at both the phylum and genus levels, with Bacillota and Bacteroidota dominant in all groups and no group-specific amplicon sequence variants detected. The same pattern held in female mice.

Allergenicity was assessed in two complementary ways. Serum immune markers, including immunoglobulin E, IgG1 and IgG2a, were measured as a preliminary screen; the only significant changes were an increase in IgE in male mice of the bovine hemoglobin group and an increase in IgG2a in male mice of the recombinant soybean hemoglobin group, both remaining within normal ranges, with no significant differences in female mice. The authors emphasize that such serum markers alone are insufficient to determine sensitization status and serve only as screening tools. The primary assessment was bioinformatic: an 80-amino-acid sliding window alignment of the soybean hemoglobin sequence against the AllergenOnline database, following the 2001 FAO/WHO expert panel recommendations and Codex Alimentarius guidelines. All matched proteins fell below the 35 percent identity threshold that suggests potential cross-reactivity risk, with the highest identities being 31.9 percent for litchi glycoside and 31.8 percent for a pollen allergen. Matches with statistically significant E-values showed extremely low identity, at 6.5 percent, and no biological relevance.

The authors conclude that recombinant soybean hemoglobin exhibited no apparent acute or subacute toxicity and no significant sensitization potential in these tests, supporting its use as a key functional ingredient to replace animal-derived hemoglobin in plant-based meat products. They also state the limits of the evidence plainly: the findings come from mouse studies over fourteen and twenty-eight days, some statistically significant changes were observed but fell within normal physiological ranges, and genotoxicity and reproductive toxicity tests remain to be investigated in future work.

Subject of Research: Safety testing of recombinant soybean hemoglobin for use in plant-based meat

Article Title: Study on the toxicity and allergenicity of recombinant soybean hemoglobin

Article References: Zhang, N., Liu, J., Hou, T., Li, J., Zhao, X., & Li, K. (2026). Study on the toxicity and allergenicity of recombinant soybean hemoglobin. Food, Nutrition and Health, 3(1), Article 39. https://doi.org/10.1007/s44403-026-00081-0

Image Credits: AI Generated

DOI: 10.1007/s44403-026-00081-0

Keywords: recombinant soybean hemoglobin, plant-based meat, food safety, toxicity testing, allergenicity, gut microbiota, BALB/c mice, hemoglobin, food-grade yeast, bioinformatics screening, serum biochemistry, sensitization