Myoadenylate Deaminase: Its Significance as a Risk Gene for Autism

Susan Costen Owens

Abstract

Aim: Myoadenylate deaminase (AMPD1) is a recognized risk gene for autism whose function is being redefined because geneticists realized that the AMPD1 phenotype may have been mischaracterized in the past.

Discussion of transitions within AMPD1 research: Originally, muscle scientists thought AMPD1 deficiency would lead to a myopathy involving pain and weakness during exercise. A world frequency of about 11% of the stop/gain defect known as c34T included many people who lacked the classically defined phenotype. Geneticists began to consider AMPD1 deficiency more as an energy problem that mainly affected the muscle. Its wider tissue expression showed AMPD1 may affect the immune system, the brain, many regulatory organs, and sleep rhythms including REM sleep.

Efforts to expand AMPD1 research: Autism researchers will need to compare people with autism who have a genetic defect on AMPD1 with people with autism who have a normal AMPD1 gene. The c34T error was first thought to produce a recessive disorder, but heterozygotes generate half the normal amount of this protein and might display a different phenotype from those who cannot make the protein at all. Additional pathogenic polymorphisms on AMPD1 may lead to a broader phenotype. The NIH is encouraging researchers to cooperate with patient advocacy groups, and such a group already exists for AMPD1.

Conclusion: Readers are invited to participate in doing research to understand the role of AMPD1 defects in autism and how different gene defects in AMPD1 will affect individual patients.

Keywords: AMPD1, autism, pleiotropy, polygenic, muscle, phenotype

1.    Aim of the article

AMPD1 is a risk gene for autism that is being reclassified by geneticists after being considered as the cause of a muscle disorder to investigators now permitting a broader phenotype that recognizes its role in the energy metabolism…an issue that could be having effects in other parts of the body. This paper will review the transition of thinking and why it is occurring and will discuss the implications for anyone assessing the significance of having a variant on AMPD1 in someone with an autism diagnosis. Readers will be encouraged to take this research further.

2. Discussion of transitions within AMPD1 research

Various genetic companies have started producing susceptibility gene panels for autism that are likely to include AMPD1. How do these companies choose which genes to include? How often do the genes they put in an autism panel have pathogenic variants that also appear outside of autism? Might the large numbers of genes in these panels make it hard for people in the autism field to understand which information will benefit patient care? This paper will make it easier to understand the significance of having a variant on AMPD1.

Research that found variants for AMPD1 in autism is why this gene began to show up on autism genetic panels (Zhang, 2015). Interest began in autism at about the same time when the function of this gene was starting to be puzzling to geneticists. As the genetics field keeps moving farther away from models where single genes lead to single conditions, AMPD1 has become a gene that will require further study using new methods that study pleiotropy and AMPD1’s possible contribution to polygenetic traits (Bellou,2020). A tool such as the NHGRI-EBI GWAS Catalog of published genome-wide association studies (Buniello et al., 2019), has ten entries for AMPD1, and three of them mention autism, but curiously, not even one of them mentions muscle.

AMPD1 was first characterized by muscle doctors, but the protein the gene makes is now understood to have broad expression outside muscle. In fact, the AMPD1 deficiency phenotype is so questioned now that the main variant which prevents the protein from being made at all is now being called a variant of unknown significance (Clinvar, 2024). What this means is that now a geneticist might avoid even calling attention to it. This paper will help explain why that response might prematurely stall research in this area.

The AMPD1 protein first appeared in the literature attributed to Fishbein et al. in 1978, (Fishbein et al.,1978) but the enzyme itself was discovered in Germany in 1928 (Schmidt, 1928).

How would the function of this gene relate to autism?

AMPD1 stands for adenosine monophosphate deaminase or myoadenylate deaminase. The “myo” stands for muscle because that is the main place this gene is expressed, but certainly muscle is not the only place in the human body the protein is found.

The Human Protein Atlas finds AMPD1 is expressed widely in the body including the cerebrum and cerebellum, the thyroid, parathyroid and adrenal glands, all kinds of immune system cells, the heart, multiple places in the respiratory system, the liver, pancreas, gall bladder, and many locations in the GI tract, the kidney, the fat and even peripheral nerves. More specifically, it is expressed in the brain in oligodendrocytes, excitatory neurons, microglial cells and in the caudate nucleus. In the immune system, this protein also functions in plasma cells, in T cells, in lymphocytes, and in the tonsils (Proteinatlas.org, n.d.). How would all these systems work differently when AMPD1 is not made at all, or when its function is compromised?

The appearance of this protein in so many places will be puzzling if you look up AMPD1 in the authoritative gene database at the National Library of Medicine and find this description:

“Adenosine monophosphate deaminase 1 catalyzes the deamination of AMP to IMP in skeletal muscle and plays an important role in the purine nucleotide cycle… Deficiency of the muscle-specific enzyme is apparently a common cause of exercise-induced myopathy and probably the most common cause of metabolic myopathy in the human” (Gene, 2024).

How often would a metabolic myopathy be recognized in a patient who has autism? Is there anything about autism that would make an AMPD1 defect produce a different phenotype?

Medline Plus furnishes more description:

“This enzyme is found in the muscles used for movement (skeletal muscles), where it plays a role in producing energy.”

They go on to explain the mechanism behind how this protein creates energy for a muscle cell:

“…the purine nucleotide cycle…reuses molecules called purines, which are a group of building blocks of DNA (nucleotides), its chemical cousin RNA, and molecules such as AMP that serve as energy sources in the cell” (Medlineplus.gov, 2024).

What their paragraph does not explain is that a lack of function in this gene will lead to elevations of adenosine after situations where the gene is normally used. After that happens, the increased adenosine will interact with adenosine receptors. Scientists have found this adenosine will create aberrations within the sleep cycle, especially in REM sleep, and REM sleep is a trait far removed from the muscles (Buyse et al., 2016).

Also, after exercise, the muscle in someone with AMPD1 deficiency will not produce ammonia normally as it does in people without this gene defect. What does that lack of ammonia change? (Norman et al., 2001) Do we have any way to recognize the tracks of that lack of ammonia?

Buyse et al. explain that this gene’s loss of function will generate an increased parasympathetic tone. They conclude that AMPD1 defects could have a profound influence on cholinergic neurotransmission and on sleep, and they argue that these facts indicate that further studies are mandatory (Buyse et al., 2016).

Does their observation furnish an unexpected context for other thoughts about cholinergic neurotransmission for autism, as expressed by Vallés and Barrantes? “Since cholinergic neurotransmission participates in the regulation of cognitive function (attention, memory, learning processes, cognitive flexibility, social interactions) brain acetylcholine receptors are likely to play an important role in the dysfunctional synapses in ASD, either directly or indirectly” (Vallés & Barrantes, 2022).

To understand that relationship further, we need to look deeper for even larger contexts.

For people involved in autism research, discussions about adenosine usually revolve around the term, “purine autism”. That quoted phrase is not a term already indexed in PubMed but it still brings up 303 articles. (PubMed, 2024) Among those are two articles published by researcher Dr. Robert Naviaux who has written two articles about purine issues in autism.

Naviaux, who led the day’s proceedings at an autism meeting sponsored by the Mitochondrial Disease Foundation in 2010, may be better known for his larger corpus of work on the cell danger response and his work in Chronic Fatigue Syndrome. One of his articles on autism explores the concept of acute hyperpurinergia in the mother. They were looking more for a cell danger signal emanating from extracellular ATP.

Extracellular ATP is not a feature that arises directly from the loss of function of AMPD1. The issue with AMPD1 instead involves either adenosine monophosphate or even adenosine itself.

The other Naviaux article explores the use of the antipurinergic drug called suramin. Suramin is a p2 antagonist, so it would block the activity of ATP at one of its receptors (Novaes et al., 2018).

Naviaux, now looking at MIA mice (developed to study neurodevelopmental and neuropsychiatric disorders), found that a single intraperitoneal dose of suramin restored normal social behavior, novelty preference and normalized 17 of 18 metabolic pathways that were disturbed in the MIA model. Most of the improvements they noted were lost after 5 weeks of drug washout, but that was a long time to watch for change. He subsequently conducted a clinical trial of this drug for children with autism. (Naviaux et al, 2018) and elaborated how their response tied in with his cell danger theory (Naviaux, 2018).

ATP’s receptors have names that begin with the letter P (Ai et al., 2023) and they seem to respond to suramin. To enable us to distinguish this activity from activity relevant to an AMPD1 defect, it is necessary to learn more about the receptors that are more specific for adenosine. Again, adenosine is the purine that is elevated with AMPD1 deficiency.

Where is the body normally expecting adenosine to come from? Gaudry et al. explain that there are many other metabolic sources of adenosine that are unrelated to AMPD1, and there are also many ways of disposing of adenosine. These complexities are nicely visualized in their Figure 1, where they show results of adenosine receptor activation: A1 activation leads to vasoconstriction, A2A and A2B lead to vasodilation, and A3 is involved with ischaemia and reperfusion (Gaudry et al, 2020).

The official gene names of these adenosine receptors are ADORA1, ADORA2A, ADORA2B and ADORA3, but in some papers these official gene names are never mentioned when speaking of these proteins (Della-Latta, 2013).

Gaudry told us why measuring adenosine’s fate becomes difficult due to its quick cellular uptake and intracellular metabolism. In fact, Gaudry et al. explained that the half-life of circulating adenosine is counted in seconds, so looking at circulating levels of this compound may underrepresent adenosine’s effects at a cellular level (Gaudry et al, 2020).

This quick disappearance of adenosine may explain why a study on Italians with autism found significant purinergic problems, but they found an elevation in inosine, rather than adenosine (Gevi et al., 2016). Conducting a similar study of people with autism who are known to have a major defect on AMPD1 would be very valuable in sorting out these differences.

In what other ways would cells producing elevated adenosine affect a patient with autism? Elevated adenosine will open the blood brain barrier (Bynoe et al., 2015). That means that any sort of toxin that is usually excluded from entry into the brain may have more access to places in the brain where it is ordinarily excluded. Raised adenosine may also trigger a physical perception of impending tissue danger, such as in ischemia, because the body is known to form extra adenosine to change things like vasodilation, and to inhibit inflammation or modify the activity of the sympathetic nervous system (Riksen et al., 2008).

Will the behavior of someone with autism and an AMPD1 defect differ after exercise? In some people, wider biochemical effects of exercise may be delayed even by hours, or even as long as the next day, but the reasons for these delays are not clear and require more research. Because of the possibility of such delays, would anyone notice a relationship of periods of exercise to behavioral change?

What about the timing of any cramping that might occur after a period of rest? Will issues like pain or cramping help identify patients that might have an AMPD1 defect?

The most common genetic finding for AMPD1 has an rsnumber of rs17602729. This snp produces what is called a stop/gain loss of function. Transcription for the protein in someone with this genetic variation will therefore fail. The decay of mRNA that follows will mean that the protein will never be made. Geneticists refer to this as ‘nonsense-mediated decay (NMD). This snp defect is frequently referred to as c34T (Norman et al., 1998).

This sort of loss of function and how often this same issue has been found in other genes has been discussed within a wider context. The AMPD1 snp was discovered to be similar to 1183 other nonsense snps. Only 8 of those snps matched variants in the OMIM database that are known to have phenotypic variations. The rarity of this occurrence led scientists to the worrisome conclusion that the biological effects of most nonsense SNPs like this one have not yet been reported (Yamaguchi-Kabata et al., 2008).

The c34T variant is mainly found in those of European descent. Its frequency is 11.7% in the largest human datasets, but a frequency of 12.6 % in those of European descent. The incidence is intermediate in Latin American populations, low in those of African descent, and extremely low in people of Asian or even South Asian descent (Snp, 2024).

Considering that data, will those with autism whose ancestors were from regions where this variant is rare have major differences in the set of other risk genes they possess? Might they possess different damaging snps on AMPD1 or other genes that also might reduce AMPD1 function, but have different inheritance patterns and perhaps different effects?

Muscle scientists called AMPD1 deficiency an autosomal recessive condition because they failed to find the phenotype they defined in people that had only one copy of c34T (Fischer et al., 2005). Even so, as early as 1994 other scientists were not finding the standard phenotype even in people who had two defective copies. That perplexing reality began casting doubt that the gene function had been characterized correctly. This is why knowing the classically-defined phenotype, and perhaps noticing its absence in someone with autism who has the c34T defect, could cast doubt that we are able to recognize other effects of AMPD1 in someone with autism (Gross, 1994).

When scientists moved instead from genotype to phenotype, they found that heterozygotes made half of what was the normal amount of this protein (Norman, 1998). That meant having one defective allele offered different challenges to those seen in homozygotes, but their situation was still far from normal. The effects from losing half function of course could be different, but also potentially damaging. With that being the case, how could this condition be called recessive?

Despite people who do not make the protein at all (homozygotes) experiencing a disruption in this major pathway, they often seemed normal in muscle energy. Might this gene defect involve a biochemical workaround in our cells that nobody has discovered? What other chemistry will change in someone with AMPD1 issues? These questions have not been answered. It is puzzling that most people with AMPD1 defects do well enough in muscle function that the defect is unknown to most muscle professionals.

Something else that should have been noticed is that the level of AMPD1 activity in a muscle will affect the ratio between fast twitch and slow twitch muscles. AMPD1 levels are highest in fast-twitch muscles (Coley et al., 2012). This difference explains why there is a preference for endurance rather than rapid speed in people with the c34T variation…a difference that may affect their choice of sport (Ginevičienė, 2018). How might this difference affect someone with autism in other ways?

Unfortunately, in the public databases that geneticists use, the old definition persists. Despite that definition, the old phenotype is so often missing in people with the proper genetics that when you look up this snp at the National Library of Medicine, it is now called a variant of unknown significance even though there are 29 articles about how it affects people and (in the past) possession of that snp gave its name to a muscle deficiency and disorder.

The significance of this change in status for AMPD1 will become even more obvious if you visit the website at the National Library of Medicine in the United States where they used data mining to associate the gene AMPD1 with what totaled to 563 conditions. Curiously, the word “muscle” is absent from their search title. What terminology did they use instead? They said they were examining variants in the gene AMPD1 that affect energy metabolism.

Does that mean the AMPD1 field should altogether switch gears and now consider AMPD1 deficiency more an energy defect? Can this relate to how you might see a patient with autism who is clearly thinking about performing an action, but just cannot summon up the drive to begin to move? Could that sort of hesitancy reflect a problem with low energy?

Looking at the datamining results, if you then search down their list of reported and associated conditions (which is much too long to report here) you will recognize many conditions that have already been found comorbid with autism (Clinvarminer, 2024). Will their long list tie AMPD1 (when it is found defective in autism) more closely to other defects that have already been associating autism to mitochondrial issues?

To what extent will having an AMPD1 defect be associated with increased comorbidities with autism? Perhaps the only way to discover that would be to find individuals with each of those other 563 conditions that were listed at the data mining site, and then to find out both their AMPD1 and their autism status.

Will we find AMPD1 comorbidity correlating more frequently or less frequently with those who did or did not have autism?

Using a similar strategy might be valid for interpreting any other genetic finding that frequently/sometimes/occasionally or even rarely is associated with autism. Scientists can first explore the genetics of people with autism that are associated with any known comorbid condition, but then they can compare those genetics to people who also have the other condition but do not have autism.

On the same data miner website, what follows next is a long list of 358 single nucleotide polymorphisms (snps) for AMPD1. 132 of those include a gnomAD frequency and 197 of the snps have rs numbers (Clinvarminer, 2024). For people unfamiliar with the gnomAD database, its usefulness is well described by an instructional video (Broad Institute, 2019) and with an excellent article by (Gudmundsson et al., 2022).

Because of the population distribution of the c34T defect, we might be tempted to think of AMPD1 as mainly affecting those descended from Europeans, but we know autism is found all over the world. Because of the former characterization of AMPD1 as primarily a muscle disorder rather than as a more global energy disorder, other genetic errors on AMPD1 may be more evenly distributed in the human population and might even be more likely to associate with body-wide energy problems, even energy issues in the brain (including what some people call brain fog) or even conditions like chronic fatigue.

Fortunately, a new article was recently written about AMPD1 looking for other pathogenic snps that have not been widely recognized. This article is not indexed in PubMed, but it may get us closer to appreciating a broader phenotype and a broader human distribution relating to genetic problems with AMPD1. We will also need to track the deleterious snps that this paper found to see if they define different phenotypes or even a different expression in human tissues.

This paper used the String database to identify 19 other genes that interact with AMPD1. How many of those genes might be identified as autism risk genes? That study also includes in Figure 7 a chart which is a KEGG Analysis of AMPD1’s metabolic pathways, and it becomes obvious that its interactions are very complex (Mohamed et al., 2023).

3. Efforts to expand AMPD1 research

The author who is writing this article began investigating AMPD1 after getting results back from three different genetics companies that identified in the author a homozygous occurrence of the c34T defect. The phenotype in the literature seemed to have no functional relevance at all at the time of testing. Curiously, after breaking an ankle accidentally, and the subsequent confinement to a wheelchair for several months, muscle issues appeared that were more like the classical AMPD1 phenotype. This was late in life for an inborn error to first appear.

The author then tried to find experts in AMPD1, but it soon became clear that neither physical therapists, doctors (including neurologists), nurses, nor people involved with professions that center around muscles, like trainers and masseuses, had ever heard of this genetic condition that should clearly affect multiple millions of people. How could AMPD1 deficiency be so unknown by professionals if it is so frequent? Is this genetic defect escaping notice by the very people who should have been recognizing and diagnosing it?

Because of the wide confusion about the phenotypes associated with AMPD1, the author started a Facebook group in 2018 for AMPD1 deficiency called AMPD1 - myoadenylate deaminase deficiency support rs17602729-AA (2024). Importantly, its membership and any access to the group is only open to those with a confirmed but recognized pathogenic genetic status on the AMPD1 gene. The reason to form the group was to see if the members could get closer to understanding what issues they had in common.

There are a hundred people still in the queue to join the group who mostly have had a muscle diagnosis from a doctor but who could not (after that diagnosis) secure adequate genetic testing. For those who knew their genetics, they reported that it was very difficult for them to find clarity about how AMPD1 relates to many of their other health issues, even including serious issues like periodic paralysis. Members also report having diagnoses that are already clearly associated with mitochondrial dysfunction. Some people in the group have a diagnosis of autism and some have family members with autism.

How can anyone with an AMPD1 defect understand the relevance of their genetics to their other issues when there is so little academic interest in this field and members can find no experts beyond the muscle doctors that they might already see?

These dilemmas are why this group is eager to see research move into defining a broader phenotype, even if the definitions include pleiotropy.

Members also report having great difficulty securing healthcare that did not make them worse. Their problems often involved pain, even outside the muscle, and they experienced problems with medications or supplements that had unpredictable and terrible effects. They talk of muscle cramping, including cramping that occurs after periods of rest.

The NIH in the United States is encouraging collaborations between patient advocacy groups and clinical investigators within a rare diseases clinical research network. That sort of coordination has to begin with establishing a common vision of what needs to be accomplished (Merkel et al., 2016).

4. Conclusion

AMPD1 is a functioning protein in humans producing a purine whose role seems to have been viewed too narrowly, leading to a failure of research to progress appropriately. The role of AMPD1 needs to be studied by people who will look more at how it functions in the energy metabolism, even looking past the muscle. This article was aimed at raising awareness of this susceptibility gene for autism and to explain to those studying autism why this defect is so poorly understood. The author appeals to readers here to become part of moving research forward so that the relevance of AMPD1 to autism can be investigated properly leading to improvements in patient care.

References

Ai, Y., Wang, H., Liu, L., Qi, Y., Tang, S., Tang, J., & Chen, N. (2023). Purine and purinergic receptors in health and disease. MedComm, 4(5), e359. https://doi.org/10.1002/mco2.359

AMPD1 - myoadenylate deaminase deficiency support rs17602729-AA (2024) In Facebook from https://www.facebook.com/groups/357418398088967

Bellou, E., Stevenson-Hoare, J., & Escott-Price, V. (2020). Polygenic risk and pleiotropy in neurodegenerative diseases. Neurobiology of disease, 142, 104953. https://doi.org/10.1016/j.nbd.2020.104953

Broad Institute; MPG Primer: gnomAD: Using large genomic data sets to interpret human genetic variation (2019) Available from: https://www.youtube.com/watch?v=J29_Pf8Ti1s

Buniello, A., MacArthur, J. A. L., Cerezo, M., Harris, L. W., Hayhurst, J., Malangone, C., McMahon, A., Morales, J., Mountjoy, E., Sollis, E., Suveges, D., Vrousgou, O., Whetzel, P. L., Amode, R., Guillen, J. A., Riat, H. S., Trevanion, S. J., Hall, P., Junkins, H., Flicek, P., … Parkinson, H. (2019). The NHGRI-EBI GWAS Catalog of published genome-wide association studies, targeted arrays and summary statistics 2019. Nucleic acids research, 47(D1), D1005–D1012. https://doi.org/10.1093/nar/gky1120

Buyse, B., Van Damme, P., Belge, C., & Testelmans, D. (2016). Possible influence of AMPD1 on cholinergic neurotransmission and sleep. Journal of sleep research, 25(1), 124–126. https://doi.org/10.1111/jsr.12341

Bynoe, M. S., Viret, C., Yan, A., & Kim, D. G. (2015). Adenosine receptor signaling: a key to opening the blood-brain door. Fluids and barriers of the CNS, 12, 20. https://doi.org/10.1186/s12987-015-0017-7

Clinvar [Internet]. Bethesda (MD): National Library of Medicine (US), National Center for Biotechnology Information; 2004 – [cited 2024 Feb 26]. Available from: https://www.ncbi.nlm.nih.gov/clinvar/variation/18271/

Clinvarminer.genetics.utah.edu [cited 2024 Feb 26] . Available from: https://clinvarminer.genetics.utah.edu/variants-by-mondo-condition/19243/gene/AMPD1

Coley, W., Rayavarapu, S., Pandey, G. S., Sabina, R. L., Van der Meulen, J. H., Ampong, B., Wortmann, R. L., Rawat, R., & Nagaraju, K. (2012). The molecular basis of skeletal muscle weakness in a mouse model of inflammatory myopathy. Arthritis and rheumatism, 64(11), 3750–3759. https://doi.org/10.1002/art.34625

Della Latta, V., Cabiati, M., Rocchiccioli, S., Del Ry, S., & Morales, M. A. (2013). The role of the adenosinergic system in lung fibrosis. Pharmacological research, 76, 182–189. https://doi.org/10.1016/j.phrs.2013.08.004

Fischer, S., Drenckhahn, C., Wolf, C., Eschrich, K., Kellermann, S., Froster, U. G., & Schober, R. (2005). Clinical significance and neuropathology of primary MADD in C34-T and G468-T mutations of the AMPD1 gene. Clinical neuropathology, 24(2), 77–85.

Fishbein WN, Armbrustmacher VW, Griffin JL. (1978). Myoadenylate deaminase deficiency: a new disease of muscle. Science; 200(4341), 545-8. https://doi: 10.1126/science.644316.

Gaudry, M., Vairo, D., Marlinge, M., Gaubert, M., Guiol, C., Mottola, G., Gariboldi, V., Deharo, P., Sadrin, S., Maixent, J. M., Fenouillet, E., Ruf, J., Guieu, R., & Paganelli, F. (2020). Adenosine and Its Receptors: An Expected Tool for the Diagnosis and Treatment of Coronary Artery and Ischemic Heart Diseases. International journal of molecular sciences, 21(15), 5321. https://doi.org/10.3390/ijms21155321

Gene [Internet]. Bethesda (MD): National Library of Medicine (US), National Center for Biotechnology Information; 2004 – [cited 2024 Feb 26]. Available from: https://www.ncbi.nlm.nih.gov/gene/270.

Gevi, F., Zolla, L., Gabriele, S., & Persico, A. M. (2016). Urinary metabolomics of young Italian autistic children supports abnormal tryptophan and purine metabolism. Molecular autism, 7, 47. https://doi.org/10.1186/s13229-016-0109-5

Gross M. (1994). New method for detection of C34-T mutation in the AMPD1 gene causing myoadenylate deaminase deficiency. Annals of the rheumatic diseases, 53(5), 353–354. https://doi.org/10.1136/ard.53.5.

Ginevičienė, V., Jakaitienė, A., Pranculis, A., Milašius, K., Tubelis, L., & Utkus, A. (2014). AMPD1 rs17602729 is associated with physical performance of sprint and power in elite Lithuanian athletes. BMC genetics, 15, 58. https://doi.org/10.1186/1471-2156-15-58

Gudmundsson, S., Singer-Berk, M., Watts, N. A., Phu, W., Goodrich, J. K., Solomonson, M., Genome Aggregation Database Consortium, Rehm, H. L., MacArthur, D. G., & O'Donnell-Luria, A. (2022). Variant interpretation using population databases: Lessons from gnomAD. Human mutation, 43(8), 1012–1030. https://doi.org/10.1002/humu.24309

Medlineplus.gov. [cited 2024 Feb 26]. Available from: https://medlineplus.gov/genetics/gene/ampd1/

Merkel, P. A., Manion, M., Gopal-Srivastava, R., Groft, S., Jinnah, H. A., Robertson, D., Krischer, J. P., & Rare Diseases Clinical Research Network (2016). The partnership of patient advocacy groups and clinical investigators in the rare diseases clinical research network. Orphanet journal of rare diseases, 11(1), 66. https://doi.org/10.1186/s13023-016-0445-8

Mohamed, A., Abdelmoneim, A, Fadl, H., Suliman, A., Mohamed, H., Abubaker, H., Elbager, S. (2023). In Silico approach for identification, prediction of AMPD1 gene nsSNPs associated with Myoadenylate Deaminase deficiency. Journal of Bioscience and Applied Research. 9(1), 17-32. 10.21608/jbaar.2023.284727.

Naviaux R. K. (2018). Antipurinergic therapy for autism-An in-depth review. Mitochondrion, 43, 1–15. https://doi.org/10.1016/j.mito.2017.12.007

Naviaux, R. K., Curtis, B., Li, K., Naviaux, J. C., Bright, A. T., Reiner, G. E., Westerfield, M., Goh, S., Alaynick, W. A., Wang, L., Capparelli, E. V., Adams, C., Sun, J., Jain, S., He, F., Arellano, D. A., Mash, L. E., Chukoskie, L., Lincoln, A., & Townsend, J. (2017). Low-dose suramin in autism spectrum disorder: a small, phase I/II, randomized clinical trial. Annals of clinical and translational neurology, 4(7), 491–505. https://doi.org/10.1002/acn3.424

Norman, B., Mahnke-Zizelman, D. K., Vallis, A., & Sabina, R. L. (1998). Genetic and other determinants of AMP deaminase activity in healthy adult skeletal muscle. Journal of applied physiology (Bethesda, Md.: 1985), 85(4), 1273–1278. https://doi.org/10.1152/jappl.1998.85.4.1273

Norman, B., Sabina, R. L., & Jansson, E. (2001). Regulation of skeletal muscle ATP catabolism by AMPD1 genotype during sprint exercise in asymptomatic subjects. Journal of applied physiology (Bethesda, Md.: 1985), 91(1), 258–264. https://doi.org/10.1152/jappl.2001.91.1.258

Novaes, R. D., Santos, E. C., Cupertino, M. C., Bastos, D. S. S., Mendonça, A. A. S., Marques-da-Silva, E. A., Cardoso, S. A., Fietto, J. L. R., & Oliveira, L. L. (2018). Purinergic Antagonist Suramin Aggravates Myocarditis and Increases Mortality by Enhancing Parasitism, Inflammation, and Reactive Tissue Damage in Trypanosoma cruzi-Infected Mice. Oxidative medicine and cellular longevity, 2018, 7385639. https://doi.org/10.1155/2018/7385639

Proteinatlas.org (n.d) Available from: https://www.proteinatlas.org/ENSG00000116748-AMPD1/tissue/primary+data and https://www.proteinatlas.org/ENSG00000116748-AMPD1/tissue and https://www.proteinatlas.org/ENSG00000116748-AMPD1/single+cell+type

PubMed [Internet] Bethesda (MD): National Library of Medicine (US), National Center for Biotechnology Information; 2004 – [cited 2024 Feb 26]. Available from: https://pubmed.ncbi.nlm.nih.gov/?term=purine+autism

Riksen, N. P., Rongen, G. A., Yellon, D., & Smits, P. (2008). Human in vivo research on the vascular effects of adenosine. European journal of pharmacology, 585(2-3), 220–227. https://doi.org/10.1016/j.ejphar.2008.01.053

Sadrin, S., Maixent, J. M., Fenouillet, E., Ruf, J., Guieu, R., & Paganelli, F. (2020). Adenosine and Its Receptors: An Expected Tool for the Diagnosis and Treatment of Coronary Artery and Ischemic Heart Diseases. International journal of molecular sciences, 21(15), 5321. https://doi.org/10.3390/ijms21155321Schmidt G. (1928) Uber fermentative desaminierung im muskel. Zeitschrift für Physikalische Chemie; 179 (243), 232.

Snp [Internet]. Bethesda (MD): National Library of Medicine (US), National Center for Biotechnology Information; 2004 – [cited 2024 Feb 26]. Available from: https://www.ncbi.nlm.nih.gov/snp/rs17602729#frequency_tab

Sollis E, Mosaku A, Abid A, Buniello A, Cerezo M, Gil L, Groza T, Güneş O, Hall P, Hayhurst J, Ibrahim A, Ji Y, John S, Lewis E, MacArthur JAL, McMahon A, Osumi-Sutherland D, Panoutsopoulou K, Pendlington Z, Ramachandran S, Stefancsik R, Stewart J, Whetzel P, Wilson R, Hindorff L, Cunningham F, Lambert SA, Inouye M, Parkinson H, Harris LW. (2022)The NHGRI-EBI GWAS Catalog: knowledgebase and deposition resource.Nucleic Acids Res.gkac1010. doi: 10.1093/nar/gkac1010 Available from: https://www.ebi.ac.uk/gwas/genes/AMPD1

Vallés, A. S., & Barrantes, F. J. (2022). Dendritic spine membrane proteome and its alterations in autistic spectrum disorder. Advances in protein chemistry and structural biology, 128, 435–474. https://doi.org/10.1016/bs.apcsb.2021.09.003

Yamaguchi-Kabata, Y., Shimada, M. K., Hayakawa, Y., Minoshima, S., Chakraborty, R., Gojobori, T., & Imanishi, T. (2008). Distribution and effects of nonsense polymorphisms in human genes. PloS one, 3(10), e3393. https://doi.org/10.1371/journal.pone.

Zhang, L., Ou, J., Xu, X., Peng, Y., Guo, H., Pan, Y., Chen, J., Wang, T., Peng, H., Liu, Q., Tian, D., Pan, Q., Zou, X., Zhao, J., Hu, Z., & Xia, K. (2015). AMPD1 functional variants associated with autism in Han Chinese population. European archives of psychiatry and clinical neuroscience, 265(6), 511–517. https://doi.org/10.1007/s00406-014-0524-6