*2.7. Statistical Methods*

Each person was used as her/his own control. In that way, individual "noise factors" could be eliminated and we were able to pairwise test whether the intra-individual responses were similar among the persons of the study. Furthermore, to achieve statistically unambiguous results all blinded participants were in the same comparison group during the first week. Pre-assessment on the sufficient group size was based on the results from our pilot tests and related volatilities. Among our relatively homogenous study group with fully comparable intra-individual measurement points, selected group size turned out to be clearly sufficient.

Statistical tests were conducted using IBM SPSS statistics software (ver. 26) and Microsoft Excel. Statistical tests for group averages of intra-individual changes were based on paired Student's *t*-tests. Unpaired t-tests were used when testing the HC and LC subgroups or the DGA and placebo groups against each other. When testing intraday 45 min changes in the DGA group, possible circadian variation was corrected by deducting the mean change in the placebo group from the DGA group observations. Squared linear correlations (R2) from scatter diagrams were calculated between selected studied variables. Technically, R<sup>2</sup> can be interpreted as the ratio of explained variation to the total variation of the explanatory variable. The significance of R<sup>2</sup> was statistically tested by the F-test. When *N* < 10, we checked the normality of the underlying data visually or using the Kolmogorov–Smirnov test. A *p*-value lower than or equal to 0.05 in a one-sided t-test was considered statistically significant (marking = \* or #) and a *p*-value lower than or equal to 0.01 as statistically very significant (marking = \*\* or ##). A non-significant *p*-value is marked by "n.s." or > 0.05, and *p* >> 0.05 in quite clear cases. All presented tests were predetermined or derived from predetermined test settings.

#### **3. Results**

### *3.1. Changes in Bilirubin and Iron Concentrations during the Entire Study Period*

Blood bilirubin concentration increased by 18.6% from Day0 to Day21 in the DGA administration group (*p* = 0.009) while in the placebo group, no statistically significant changes were noticed (Figure 1A). Furthermore, the Day21 changes from Day0 in the DGA group differed statistically significantly (*p* = 0.023) from the respective changes in the placebo group (Figure 1A). There was an interesting decline by 8.8% in blood Fe from Day0 to Day4 (*p* = 0.042). Blood bilirubin and Fe concentration shared a similar pattern during the study period (Figure 1A,B).

**Figure 1.** End products of the HO-1 pathway in blood 12 h after the last DGA or placebo dose: (**A**) Mean (±SEM) blood bilirubin concentration, (**B**) Mean (±SEM) blood Fe concentration. **Notes:** (1) The Day21 DGA and Placebo averages are indexed to Day0 so that Day21 fully reflects the change from Day0. (2) \* and \*\* indicate paired *t*-test, # indicates non-paired *t*-test between the%-changes from Day0 to Day21.

#### *3.2. Biomarkers Downstream of Bilirubin and the HO-1 Pathway during the Entire Study Period*

Figure 2A shows that the relative changes between GlycA and bilirubin from Day0 to Day21 correlated significantly (R2 = 0.353, *p* = 0.015). Respectively, the changes between triglycerides and bilirubin correlated statistically significantly (R2 = 0.305, *p* = 0.032) (Figure 2B) and a not-significant correlation was noticed between the changes Day4 to Day21 in IL-6 and bilirubin (Figure 2C).

**Figure 2.** Bilirubin associated Day21 anti-inflammatory, fat metabolic, and mRNA expression effects: (**A**) scatter diagram and R2 of the percent changes from Day0 to Day21 between GlycA and bilirubin, (**B**) scatter diagram and R<sup>2</sup> of the percent changes from Day0 to Day21 between blood triglycerides and bilirubin, (**C**) scatter diagram and R2 of the percent changes from Day4 to Day21 between IL-6 and bilirubin, (**D**) mean (±SEM) blood IL-6 concentration, (**E**) mean (±SEM) mRNA expression of HIF-1α from collected WBCs. **Notes:** (1) In (**D**,**E**), the Day21 DGA and placebo averages are indexed to Day0 so that Day21 fully reflects the change from Day0. (2) \* and \*\* indicate paired *t*-test, # indicates non-paired *t*-test between the %-changes from Day0 to Day21. In (**A**–**C**), the *p*-values are based on an F-test. (3) One Day21 observation was deducted from (**B**) as an outlier. Additionally, one Day0 IL-6 observation was excluded from (**D**) when comparing the changes from Day0 to Day21 in the placebo group against the DGA group.

Strong upregulation of blood bilirubin from Day4 to Day21 was typically accompanied by a reduction of subclinical inflammation (IL-6) at the individual level (Figure 2C). IL-6 declined by 19.0% from Day0 to Day21 in the DGA administration group (*p* = 0.002), while in the placebo group it declined only by 2.6% (Figure 2D). Furthermore, the individual changes of IL-6 in the DGA compared to the placebo group differed statistically significantly (Figure 2D) [27].

HIF-1α mRNA expression in WBCs (Figure 2E) shared a similar pattern in all observation points to blood bilirubin in the DGA group (Figure 1A). HIF-1α mRNA expression increased by 37.0% from Day0 to Day21 in the DGA administration group (*p* = 0.014) while in the placebo group there was no change (Figure 2D). Furthermore, the 49% increase (*p* = 0.004) in HIF-1α mRNA expression resembles the 24% increase in blood bilirubin from Day4 to Day21.

HO-1 mRNA expression in WBCs was very weak on Day0. Eight of the successfully analyzed twenty-five WBC samples showed a lower signal than the 0.29 detection limit and were set to zero (graph not shown). On Day4, the number of zeros increased to 11, which may suggest that the HO-1 mRNA expression was downregulated during the 4-day DGA regimen. Further in line with blood bilirubin (Figure 1A), average HO-1 mRNA expression increased in the DGA group by 34.0% on Day21 from Day0. However, poor mRNA signals destroyed the possibility of drawing statistical conclusions on the changes of HO-1 mRNA expression in WBCs.

#### *3.3. Acute 45 Min Activation of the HO-1 Pathway on Day4*

Blood bilirubin concentration rose by 2.3% 45 min after the acute dose in the DGA group (*p* = 0.066) but not in the placebo group (Figure 3A). Blood Fe rose by 9.1% (*p* < 0.001) in the DGA group but not in the placebo group (Figure 3B). Furthermore, there existed a statistically very significant positive correlation between the 45-min %-changes of Fe and bilirubin at the intra-individual level in the DGA group (Figure 3C).

**Figure 3.** HO-1 pathway end products excluding CO in blood 45 min after acute DGA dose: (**A**) 45 min average intraindividual percent change in bilirubin (±SEM), (**B**) similar 45 min change in Fe (±SEM), (**C**) scatter diagram between 45 min %-changes of bilirubin and Fe after acute DGA dose. **Notes:** \*\* indicate paired *t*-test, # indicates non-paired *t*-test between the %-changes in 45 min. In (**C**), the *p*-value is based on an F-test.

The average relative increase in blood bilirubin (Figure 3A) was much lower than in Fe (Figure 3B) as some of the individual relative changes in blood bilirubin concentrations were negative, unlike in Fe (Figure 3C).

#### *3.4. 4-Day Changes of Blood Bilirubin Are Tightly Linked to the Changes of Its Carrier in Blood*

On Day0, bilirubin concentration in the HC subgroup was on average 12.8 μmol/L, while in the LC subgroup it was 9.6 μM (*p* = 0.085). Higher bilirubin in the HC compared

to the LC subgroup remained throughout the 4-day DGA regimen until Day4 (Figure 4A). Conversely, Day0 blood albumin concentration was very similar in the HC and LC subgroups (Figure 4B). Interestingly, both blood albumin and bilirubin declined in the LC subgroup during the 4-day DGA regimen. Moreover, there existed a very tight association (*p* < 0.001) at an individual level between molar *changes* of albumin and bilirubin during the 4-day DGA regimen (Figure 4C). In the HC subgroup, there was no correlation between the changes of albumin and bilirubin during the 4-day DGA regimen (graph not shown). Furthermore, there was no association between the changes of Fe and albumin (R2 = 0.017, *p* >> 0.05) in the LC subgroup. We additionally measured the association between the 4-day changes of blood free fatty acids (= total fatty acids minus TGs) and blood albumin, and found only a mild positive correlation (R2 = 0.075, *p* > 0.05, graph not shown).

**Figure 4.** 4-day DGA regimen activates hepatic bilirubin metabolism and Fe storage in the LC subgroup: (**A**) mean (±SEM) blood bilirubin in the HC and LC subgroups separately, (**B**) mean (±SEM) blood albumin in HC and LC subgroups separately, (**C**) cross correlation between intraindividual 4-day molar changes of albumin (vertical scale) and bilirubin (horizontal scale) in the LC subgroup. (**D**) mean (±SEM) blood ALP in the HC and LC subgroups separately, (**E**) mean (±SEM) blood Fe in the HC and LC subgroups separately, and (**D**) cross correlation between intra-individual four-day percent changes of Fe (vertical scale) and bilirubin (horizontal scale) in the whole group. **Notes:** \* and \*\* indicate paired *t*-test from Day0 to Day4, # indicates non-paired *t*-test between the bilirubin concentration on Day4. In (**C**,**F**), the *p*-values are based on an F-test.

Remarkably, plasma alkaline phosphatase (ALP) was reduced in each LC participant during the 4-day DGA regimen (*p* < 0.0001, Figure 4D). Also, blood Fe concentration was reduced statistically very significantly in the LC subgroup (Figure 4E). The reduction of average blood Fe was 16.3% in the LC subgroup. Finally, there was a very strong correlation between the 4-day relative changes of blood Fe and bilirubin (Figure 4F, *p* < 0.0001).

## *3.5. 4-Day Responses of Blood Bilirubin and Fe Concentrations Are "Inversely" Associated with the Changes of Blood TGs and Subclinical Inflammation (GlycA)*

Blood TGs concentration deviated strongly between the HC and LC subgroups at Day0 (Figure 5A). During the 4-day DGA regimen, the difference in TGs between the HC and LC subgroups reduced. Especially in the LC subgroup, there was a significant reduction in blood TGs. The Day0 concentrations in the HC and LC subgroups were opposite compared to each other in blood bilirubin (Figure 4A) and TGs. Nevertheless, the responses in bilirubin and TGs to the 4-day DGA regimen were in the same directions (Figures 4A and 5A). Moreover, there existed a statistically significant positive correlation between the 4-day individual percent changes of blood TGs and blood bilirubin (Figure 5B), and with the 4-day percent changes of blood Fe (Figure 5C).

**Figure 5.** 4-day DGA regimen activates the LC subgroup: (**A**) mean (±SEM) blood TGs in the HC and LC subgroups separately, (**B**,**C**) cross correlation between intra-individual four-day percent changes of TGs (vertical scale) and (**B**) blood bilirubin (horizontal scale) and (**C**) blood Fe (horizontal scale) in the whole group. (**D**) Mean (±SEM) blood GlycA in the HC and LC subgroups separately. (**E,F**) Cross correlation between intraindividual four-day percent changes of GlycA (vertical scale) and € blood bilirubin (horizontal scale) and (**F**) blood Fe (horizontal scale) in the whole group. **Notes:** \* and \*\* indicate paired *t*-test from Day0 to Day4, # indicates non-paired *t*-test between the TGs and GlycA concentrations on Day0. In (**B**,**C**) and (**E**,**F**), the *p*-values are based on an F-test.

Blood subclinical inflammation (GlycA concentration) deviated strongly between the HC and LC subgroups at Day0 (Figure 5D). During the 4-day DGA regimen, the difference reduced and, especially in the LC subgroup, there occurred a remarkable reduction in blood GlycA (Figure 5D). There existed a statistically significant positive correlation at an individual level between the 4-day relative changes of GlycA and bilirubin (Figure 5E). Moreover, the 4-day relative changes of GlycA correlated strongly with the relative changes

of Fe (Figure 5F). Finally, responses in GlycA during the 4-day DGA regimen resemble the respective changes in TGs (Figure 5A,D). There existed a very strong association between the 4-day relative changes of blood TGs and GlycA (R2 = 0.686, *p* < 0.001, graph not shown).
