Contribution of ROS and metabolic status to neonatal and adult CD8+ T cell activation

In neonatal T cells, a low response to infection contributes to a high incidence of morbidity and mortality of neonates. Here we have evaluated the impact of the cytoplasmic and mitochondrial levels of Reactive Oxygen Species of adult and neonatal CD8+ T cells on their activation potential. We have also constructed a logical model connecting metabolism and ROS with T cell signaling. Our model indicates the interplay between antigen recognition, ROS and metabolic status in T cell responses. This model displays alternative stable states corresponding to different cell fates, i.e. quiescent, activated and anergic states, depending on ROS levels. Stochastic simulations with this model further indicate that differences in ROS status at the cell population level contribute to the lower activation rate of neonatal, compared to adult, CD8+ T cells upon TCR engagement. These results are relevant for neonatal health care. Our model can serve to analyze the impact of metabolic shift during cancer in which, similar to neonatal cells, a high glycolytic rate and low concentrations of glutamine and arginine promote tumor tolerance.

1. Important details are missing in the methods, particularly the number of cord blood and adult samples obtained. Were these from particular sexes? What was the time from the obtaining of the cord blood samples to their processing? Was this variable and did this affect the result. What were the age and sex of the adult donors? Was there a similar interval between taking of blood from adult donors and its processing? Was the cord blood from natural births or caesarians. Had the mothers received any drugs during labour that might have affected the T cell responses. Was any attempt made to obtain parallel blood from the mothers to compare to the cord blood as this would exclude any drug or hormonal confounders that might be affecting the cord blood T cells.
For each comparison the minimum number of samples included in the neonate vs adult groups was three. For most cases the number of samples was five. The total number of samples used for this study was 30 samples for neonatal CD8 + T cells and 30 samples of adult cells. Because we were using purified naïve cells, only 3 to 6 million CD8 + T cells were obtained from cord blood, and 5 to 10 million from adult samples. No eligibility criteria were defined regarding the donors' gender, this parameter was distributed randomly across the samples, and an equivalent number of male and female donors was used for neonatal and adult cells. No differences related to gender were detected. For adult samples, the ages of the donors were in the range between 26 and 40 years. All cord blood samples were obtained from vaginal deliveries, no drugs known to affect T cells responses were administered to the mothers during labor, mostly because of budget limitations in the hospital. All the samples were immediately processed after collection. This information has been included in the Materials and Methods section of the revised manuscript, lines: 130-150. We did not collect blood from mothers because we only obtained permission for noninvasive procedures.

Fig 2 seems to suggest there were only 3 cord blood and 3 adult samples obtained?
How do the authors know these responses are representative? As cord blood may not be representative of blood of young children due to prevailing maternal and birth stress factors, was any attempt made to obtain blood from young babies to show that these exhibited the same pattern as the cord blood samples?
In Fig 2A, we obtained 5 adult cells samples and 5 neonatal cell samples, in Figs 2B and 2C we obtained only 3 adult samples and 3 neonatal samples. For Fig 2D, we captured 4 neonatal blood samples. We made more evaluations of CD69 protein with the same results. Other members in the lab have also measured this protein with similar results, but together with ROS measurements, we made 3 evaluations with statistically significant results. The data shown in Fig 2 are consistent with a high ROS levels in the neonatal cells and come from independent biological samples (because of the limitations on the number of cells), thus strengthening our conclusions. Mothers in Hospital Parres do not receive anesthetics, hormones or drugs during delivery, unless they go into caesariansection. All samples were captured from vaginal deliveries, with no treatment whatsoever. The number of CD8 + T cells for about 40 ml of cord blood is about 3 and 5 million cells. It is almost impossible to obtain CD8+T cells from infants later in life. In addition, the amount of blood that could be authorized for young infants' samples is 1 mL. Care was taken to obtain the blood immediately after delivery and before placenta expulsion, to have a sample more representative of the infant blood. Cells were put in culture medium during the purification period of two days, which arrests the cells in resting conditions. We have compared cord blood from vaginal deliveries versus programmed caesarian sections and these proved to be largely similar. Although published reports claim that cord blood is not representative of the infant blood, these do not state how they obtained the samples (Olin et al. (2018). Cell 174 (5): 1277-92). Note that most labs are only allowed to get cord blood samples from arterial blood of expelled placentas, in which the blood started the coagulation process and both serum and cell types are altered because they get trapped in the clot. In contrast, in our study, cord blood was sampled immediately after delivery and before placenta expulsion.

In view of the above the authors need to add a paragraph at the end of the study describing shortcomings in the study design and cautions in interpretation of the data.
This has been included in the revised manuscript, lines 510-561.

Reviewer # 2: The overall results of the study are quite interesting with respect to the mROS differences. There are a few issues that should be addressed to improve the manuscript:
1. A through reading of the manuscript to improve the wording would help the reader. There are few word choices and sentence mechanics all throughout the manuscript that make it a bit difficult to read.
The revised manuscript has been reviewed by Prof. Christopher Ian Pogson, English born scientist, graduated in Cambridge, UK. Former head of the Biochemistry Department of Manchester University and Editor of The Biochemical Journal now retired. figure 2, the authors found a large difference in HV-1 expression between infants and adults. However, it appears that GADPH was used as the normalization control which is no longer appropriate. GADPH is so widely impacted by immunological processes (i.e. TNFa or IFNg mRNA transcripts bind to it) that it is not an appropriate housekeeping control. Their data may still be relevant, but they should ensure the rigor of their findings with another calibrator, especially when comparing neonates to adults where transcripts such as IFNg are widely different during activation.

In
A set of experiments were performed with the aim of selecting the proper reference gene for qPCR comparisons. We are enclosing below a table with information regarding the quality of the primers.