Thyroid hormones affect a wide variety of metabolic and physiological processes, and were recently linked to increased spleen size as an apparent adaptation to breath-hold diving (Ilardo et al., 2018). However, the spleen-thyroid relationship is not new; as early as 1,927 hyperthyroidism was reported in medical contexts to be associated with large spleen size (Baldridge and Peterson, 1927), and at one time splenomegaly was used as a diagnostic tool. The exact mechanisms by which thyroid hormones may affect spleen size are yet unknown. Here, we report an erythropoietin-independent mechanism of red blood cell (RBC) production as a consequence of thyroid hormone stimulation via the pharmacological inactivation of Phosphodiesterase 10A (PDE10A).

Phosphodiesterase 10A is a phosphodiesterase capable of hydrolyzing both cAMP and cGMP, thus regulating cyclic nucleotide signaling (Fujishige et al., 1999). It was recently discovered that PDE10A expression has been genetically reduced in a population called the Bajau as a consequence of natural selection acting on the ability to sustain repeated bouts of apneic diving (Ilardo et al., 2018). This reduction in expression has resulted in two apparent phenotypes, the first of which is increased thyroid hormone levels. Although thyroid hormones have not been directly measured in the Bajau population, a genome wide association scan (GWAS) for Thyroid Stimulating Hormone (TSH) levels (a proxy for thyroid hormones with high sensitivity), revealed a strong association between SNPs in PDE10A as well as in related phosphodiesterase PDE8B, which encodes a high affinity cAMP-specific phosphodiesterase (Arnaud-Lopez et al., 2008; Soto-Pedre et al., 2017). The second phenotype associated with reduced PDE10A expression is splenomegaly. Spleen size has been proposed to affect dive capacity owing to the spleen’s role in the mammalian dive reflex, in which it contracts during diving in order to expel a bolus of oxygenated RBCs thus expanding available oxygen (Thornton et al., 2001). The large spleen phenotype observed in the Bajau has been proposed to enhance their dive capacity, though the mechanisms underlying the adaptations are unknown.

Here we use a pharmacological approach to inhibit PDE10A in mice in order to investigate the molecular mechanisms by which thyroid hormones are connected to an increase in spleen size in the Bajau. We utilized the highly selective PDE10A inhibitor MP-TRA 10 (MP-10), which has previously been shown to potentiate thermogenesis, suggesting a thermoregulatory role (Hankir et al., 2016). Through daily injections, we were able to mimic the genetic effect of the Bajau adaptation in mice. We subsequently observed a significant increase in spleen size compared to control animals, as well as a corresponding increase in RBC count originating in an excess proliferation of erythrocytic precursors in the spleen, which in mice, unlike in humans, is a site of erythropoiesis. Further, we demonstrate that this stimulation of proliferation is likely erythropoietin (EPO) independent, suggesting a direct role of thyroid hormones on erythrocytic precursor development. These data suggest the benefits of Bajau PDE10A adaptation are two-fold; first, the increased RBC count would increase oxygen capacity allowing for longer dives on a single breath, and secondly the corresponding increase in spleen size creates a larger reservoir in which the oxygenated red blood cells may be stored.

We chose to divide our pharmacological model into two studies conducted over different time courses; a long-term study with lower dosage to represent chronic, mild inhibition of PDE10A, and a short-term study with higher dosage to represent more acute inhibition. For both studies there is precedent in the literature (Hankir et al., 2016), and both provide different perspectives to inform us of the mechanisms of response to PDE10A inhibition. While the long-term study may more closely represent inhibition in humans with the Bajau variant in PDE10A, the short term study may more closely resemble a knockout model. For our long-term study, we injected 4-week old male mice daily i.p. with a low (10 mg/kg) dosage of MP-10 over a period of 9 weeks. We observed lower body weights in treated mice compared to control mice (Figure 1). These differences are consistent with previous observations in mouse studies that administered similar dosages of MP-10 (Hankir et al., 2016), and we believe are likely a consequence of increased basal metabolic rate (BMR) resulting from thyroid hormone stimulation. Following 0, 4, and 9 weeks of treatment, we measured spleen size in vivo using a small animal ultrasound. We found that after 4 and 9 weeks of treatment the treated mice had larger spleens in vivo than the control animals (p = 0.0634, and p = 0.0374, respectively), consistent with the large spleen phenotype observed in the Bajau (Figure 1; Ilardo et al., 2018). We repeated the experiment at a higher dosage (30 mg/kg) for a shorter time course of 1 week. Following 1 week of treatment, treated mice had gained significantly less weight than controls (p = 0.0019) (Figure 2). Through ex vivo spleen measurements of length and width, we determined that the treated mice again had significantly larger spleens than the control mice (p = 0.0059, Figure 2). However, the spleens of treated mice were also significantly less dense than those of the control animals (p < 0.0001, Figure 2), as calculated by a ratio of measured splenic area and splenic mass. Spleen sections were stained with H&E, and differences were clearly visible and appear to arise from increased distance between cells (Figure 3). We also observed that the treated mice had higher RBC count, hematocrit (Hct), and hemoglobin (Hgb) (Figure 4). While our hematological values were consistently low across all animals, they were within the normal range for young animals (Wozniak et al., 2021). These findings are consistent with the spleen becoming stretched, and thus less dense, suggesting that they may serve as a reservoir for a greater number of red blood cells. It is, however, difficult to say whether this increase is indeed due to a greater RBC mass because of the potentially confounding effect of changes in blood volume. Such a determination would require chromium labeling or CO rebreathing to accurately determine RBC mass. Using GWAS data from the UK Biobank (accessed using the Global Biobank Engine, Stanford, CA,¹ December 2019), we identified consistent results in humans of European ancestry. Humans carrying the Bajau variant at previously identified SNPs known to influence PDE10A expression were also found to have significantly higher RBC count than those carrying the ancestral variant (Table 1). We have not, however, confirmed this association in the Bajau population.

Through flow cytometry, we were able to attribute the difference in hematological measures to differences in erythrocytic precursor populations (Figure 5). We found a significant increase in proliferation at nearly every stage of cell proliferation (Figure 5 and Table 2 for a summary of p-values), however late stages (III and IV, representing late basophilic erythroblasts and orthochromatophillic erythroblasts, respectively) were also found to have a significant increase in apoptosis, resulting in an overall lower proportional abundance of cells in these stages. This balancing effect may be mediated by thyroid hormones and serve to ensure a normal number of reticulocytes is released.

In order to investigate whether the underlying cause of increased RBC precursor proliferation could be thyroid hormone induced stimulation of erythropoietin (Epo), we used qPCR to measure Epo mRNA in kidney. We found no difference in Epo expression between control and MP-10 treated mice (p = 0.48). Using GWAS data from Prevention of REnal and Vascular ENd (PREVEND), we confirmed this result in humans of European ancestry. In spite of a clear, significant association with thyroid hormone levels previously reported in Ilardo et al. (2018), there is no association between the Bajau PDE10A SNPs and Epo levels (Table 1). We investigated additional SNPs in other genes known to be strongly and significantly associated with thyroid hormone levels in European populations, none of which were found to be associated with Epo levels (Table 3). We have yet to confirm these results in Bajau individuals.

Through a pharmacological approach in mice, we were able to simulate the Bajau large spleen adaptation to breath hold diving. We administered daily injections of the selective PDE10A inhibitor MP-10 daily via ip injection at low (to measure chronic inhibition) and high (to measure acute inhibition) dosages for 9 weeks and 1 week, respectively. Using a small-animal ultrasound, we determined that the treated mice had significantly larger spleens in vivo than the control animals, confirming the phenotype observed in the Bajau. This phenotype was replicable ex vivo, however the spleens of treated mice were significantly less dense than those of the control animals. We attribute this difference in splenic density to a stretching of the tissue to accommodate a greater quantity of RBCs, apparent through hematological testing.

We attribute the increase in RBC count to an observed significant excess of early stage erythrocytic precursors in treated mice, which has been previously identified as a response to thyroid hormone stimulation (Angelin-Duclos et al., 2005). Thyroid hormones are known to be required for terminal erythroid differentiation, however the exact molecular mechanisms underlying thyroid hormone function on erythropoiesis are unknown (Gao et al., 2017). In agreement with our results, thyroid hormones have been previously demonstrated to directly affect proliferation of erythroid progenitors (Golde et al., 1977; Popovic et al., 1977; Dainiak et al., 1978; van Gucht et al., 2017). This may be attributable to higher levels of TH nuclear receptors α (TRα) in early progenitors compared to late stage erythroblasts, as has been previously demonstrated to underlie variable responses to thyroid hormones at different stages of development (Gao et al., 2017).

We also note a significant increase in apoptosis of late-stage erythrocytic precursors in the spleens of treated mice, a previously characterized response to splenomegaly (Ilesanmi, 2010). We suggest that thyroid hormones mediate an apoptotic feedback mechanism that, in response to increased erythrocytic precursor stimulation, controls the number of reticulocytes that are released and is preventative of red blood cell excess. However, the apparent apoptotic mechanism proves insufficient in the short term to overcome the excess production of early precursors; after 1 week of treatment, the experimental mice displayed pronounced polycythemia (Figure 4). It appears that over longer periods, in the case of our study a 9-week administration of MP-10, the feedback returns hematological values to normal levels (Figure 4). This could be a limitation of the pharmacological model and the result of slight inconsistencies in delivery of the low dosage of MP-10 over the 9 weeks. A knockout model would provide more consistent inhibition of PDE10A and more reliable hematological values. It could, however, be truly representative of what occurs in humans and therefore the reason hyperthyroidism has not yet been consistently linked to polycythemia.

A reduction in RBC count, and a corresponding reduction in EPO has been previously reported in hypothyroid patients, however the reduction in EPO was thought to be associated with reduced oxygen requirement due to diminished BMR (Das et al., 1975). In fact, it has been suggested that thyroid hormones only induce EPO production under hypoxic conditions (Fandrey et al., 1994). In vitro studies have shown that thyroid hormones can stimulate erythropoiesis directly in systems where hormonal effects on EPO production are eliminated (Malgor et al., 1975; Golde et al., 1977), thus suggesting an EPO independent mechanism, consistent with our findings which showed no difference in Epo expression between treated and control animals. It may be that slight differences in expression are obscured by the nature of the measurement. Unfortunately, we were not able to measure TH or EPO in the animals directly. In addition, it is important to note that while Gapdh is frequently used for normalization in other TH studies, it is a HIF-dependent gene and as such may change in expression along with our gene of interest.

It is important to note that the adult human spleen is not a site of erythropoiesis as it is in mice. Because of this, observed cell population differences in the mouse spleen should be considered analogous to bone marrow in humans (Brodsky et al., 1966; Bresnick et al., 2018). In addition, erythroblastic islands, which are essential for the maturation of erythroblasts that are fated to enucleate, have been demonstrated in vivo in fetal splenic red pulp (Manwani and Bieker, 2008). Therefore, the observations in this study may be extrapolated to human fetal development and the observed changes in spleen size may occur developmentally.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

The animal study was reviewed and approved by the University of Utah IACUC.

MI devised and performed the experiments and wrote the manuscript. MS assisted with the designing and implementing all the experiments as well as data interpretation. NG, NV, PiV, and PeV generated, analyzed, and interpreted the EPO GWAS data. MR assisted with the experiments. EL assisted in the experimental design. All authors reviewed the manuscript.

MI was employed by the company Maze Therapeutics. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.