On 2016 Feb 18, Abdulla A-B Badawy commented:

None

On 2016 Feb 29, Simon Young commented:

In the J Psychopharmacol article under discussion Badawy and Dougherty “concluded that the ATD formulation of Young et al. (1985) may involve not only serotonin depletion, but also that of DA and possibly also NA, because it also induces a decrease in the [Phe + Tyr]/[BCAA + Trp] ratio, thereby decreasing Tyr availability to the brain.” Their concern was that “This could impact catecholamine-specific behaviours and those influenced by catecholamines”. They now argue that brain tyrosine may be increased, rather then decreased as argued in the article, because while some studies find a decrease in the ratio others find an increase, and because CSF tyrosine shows small increases at later times after administration of the 100g ATD mixture, when CSF is sampled for 12 hours Carpenter LL, 1998. This raises two important issues.

(a) The fact that the ratio increases in some circumstances and decreases in others supports my contention that relatively small changes in the ratio do not provide useful information.

(b) Badawy and Dougherty fail to respond to the fact that CSF HVA does not change after administration of the ATD mixture Carpenter LL, 1998, indicating that synthesis of dopamine (and hence synthesis of its metabolic product norepinephrine) is not altered. The small increase in tyrosine with no increase in dopamine synthesis is not surprising as human brain tryptophan hydroxylase is regulated not only by substrate availability but also by product inhibition and phosphorylation Almås B, 1992.

Badawy and Dougherty agree that protein intake can alter cognition, and presumably would agree that the “whey protein fraction glycomacropeptide” that they suggest using in ATD studies also may influence cognition. They do not respond to the point I made that normal physiological mechanisms (e.g. food intake and diurnal rhythms) can alter cognition. Therefore, the issue is not whether cognition will change over time after the administration of a control amino acid mixture, but whether such a change is outside the normal physiological range. As long as an ATD mixture and its control mixture differ only in tryptophan levels any change mediated by factors other than changes in tryptophan will occur after control and ATD treatments. Therefore, the difference in the effects of the two mixtures will be due only to changes in tryptophan. Badawy and Dougherty suggest “that it is prudent to use a control formulation that exerts a zero effect on all ratios”. As discussed above I do not feel that small changes in ratios are reliable indicators of anything occurring in the brain, an assertion not disputed by Badawy and Dougherty. Even if one did believe in the validity of small changes in ratios, the discussion above illustrates that ratios can vary from one study to another (both between different labs and also within labs). Therefore, Badawy and Dougherty’s suggestion “to use a control formulation that exerts a zero effect on all ratios” is probably not realistic.

I still believe that the criticisms of Badawy and Dougherty of the 100g ATD mixture are not based on experimental evidence.

On 2016 Feb 18, Abdulla A-B Badawy commented:

We thank Young for his second response and agree with him on many of his statements. We also confirm in agreement with Young that our major concern is the specificity of his control formulation for ATD (and also that of the control formulation for ATPD). Young quotes the increase in CSF [Tyr] reported by Carpenter LL, 1998 as evidence against a decrease in the Tyr ratio after ATD. Others have in fact reported increased Tyr ratios after ATD (see Booij L, 2005) and we have never questioned these findings in studies using the ATD test formulations. The data reported by Carpenter LL, 1998 Booij L, 2005 demonstrate that the ATD test formulation does increase Tyr availability to the brain. In our opinion, this confirms the lack of specificity of the ATD formulation for serotonin. We now propose further that the Tyr and Phe contents of the ATD formulation of Young SN, 1985 should be decreased. The level of such a decrease should be determined in conjunction with our proposed lowering of the BCAA content. To prevent a decrease in Tyr availability to the brain after acute Trp loading (ATL), investigators will need to optimise the ATL formulation by titrating levels of Trp, Tyr and Phe along with those of BCAA.

Regarding the control formulation, our point is that, while decreases in ratios of 30-40% may not influence behaviours such as mood and cognition in normal subjects, they may do so in subjects at risk, e.g. depressed patients, who already have a ~ 30% lower [Trp] and [Trp]/[CAA] ratio (Badawy AA, 2013). Thus, a low-dose ATD (25g) given to remitted depressed patients is associated with altered cognitive function, but not mood (Booij L, 2005; Haddad AD, 2009) and it was suggested (Booij L, 2005) that the cognitive changes observed are more sensitive markers of 5-HT function than symptoms, unless different mechanisms apply. We maintain our position that it is prudent to use a control formulation that exerts a zero effect on all ratios and this is best achieved by lowering the BCAA content from the usual 30% to 18%.

Our suggestion of lowering the BCAA content of both the control and test formulations, which can, in addition to normalising the relevant ratios, minimise the effects of BCAA on behaviour, including cognitive function, is not without support. BCAA are nitrogen donors for the synthesis of glutamate and GABA and can thus influence cognition by modulating glutamatergic function (see references in Badawy AA, 2015). Enhanced cognition by BCAA in athletes has been widely reported. Dietary BCAA ameliorate injury-induced cognitive impairment (Cole JT, 2010). Young rightly states that food intake can also enhance memory and it is interesting in this regard that cognition is enhanced by a high protein meal, which has been attributed to increased levels of BCAA and Phe (Jakobsen LH, 2011). Objective cognitive function is variously influenced acutely by the balance between carbohydrate and protein within a few hours following food intake, with cognitive performance being enhanced by a high protein or a balanced meal, but not by a high carbohydrate meal (Fischer K, 2002). These latter authors also attributed this enhancement to changes in LNAA ratios almost certainly involving the release of BCAA from proteins. In our opinion, enhancing cognition by the BCAA in the ATD control and test formulations could impact mood changes and it is therefore desirable to minimise this confounder in ATD studies by lowering the BCAA content.

We, and no doubt also Young, hope that our exchanges will stimulate the psychopharmacological research community to consider the various issues discussed in our comments in future studies.

On 2016 Feb 08, Simon Young commented:

I thank Badawy and Dougherty for their comments. The main points of disagreement concern the specificity of the 100g ATD mixture, and whether the relevant control mixture might lower serotonin synthesis and therefore mood, cognition and behavior. The arguments that Badawy and Dougherty make on both points depend on the plasma ratio of the relevant amino acid precursors to the LNAA. They suggest that the 100g ATD mixture lowers dopamine synthesis, as the ratio of Tyr or Phe + Tyr to the LNAA, while varying greatly between different studies, declines by up to 60%. They do not comment on the study I cited in my original comment showing that CSF tyrosine remained unchanged for the first few hours after the amino acid mixture, but rose slightly between about 5 and 10 hours after the mixture Carpenter LL, 1998. CSF levels of the dopamine metabolite homovanillic acid (HVA) remained unchanged. Measurement of biogenic amine precursors and metabolites has been used to study human CNS metabolism for over 50 years, and while it is not a perfect method it is generally accepted as a valid method by the scientific community, as indicated by the huge number of articles using this technique in the literature. This raises the question why the conclusion from plasma amino acid ratios differs from the CSF study concerning the specificity of the 100g ATD mixture.

The ratio of Trp or Tyr or Phe + Tyr to the LNAA is known to be only an approximation of the rate of transport of an amino acid across the blood-brain barrier, as it does not take into account the different affinities of the different amino acids for the transporter. This can be done for rats using a calculation based on Michaelis-Menten kinetics Smith QR, 1987, but it cannot be done for humans as the affinities of the different amino acids for the transporter in human brain are not know. Badawy and Dougherty point out that there is a correlation between CSF Trp and the plasma [Trp]/[LNAA] ratio. However, after ATD the Pearson’s correlation was r = 0.41 Moreno FA, 2010. R2 was 0.168, indicating that less than 20% of the variance in the plasma [Trp]/[LNAA] ratio is related to CSF Trp values. The inaccuracy of the plasma ratios as a measure of changes in brain amino acids explains why they suggest that the 100g ATD mixture causes a decline in brain tyrosine and dopamine synthesis, while the CSF measures indicate that this is not so. While the ratios of Trp, Tyr, or Phe + Tyr to the LNAA may be useful to demonstrate that ATD or APTD was effective in depleting precursor levels in any participant, where the effects are very large, these plasma measurements do not provide reliable information when the changes are smaller.

Badawy and Dougherty suggest that the control mixture should have a lower level of the branched chain amino acids (BCAA) than the control mixture used with the 100g ATD mixture because (i) this will decrease the lowering of the [Trp]/[LNAA] ratio, and (ii) reducing the BCAA content “can also minimise the BCAA effects on cognition and hence potential modulation of behavioural outcomes”. While I am not aware of any evidence for BCAA influencing behavior in humans, I suggest that all control mixtures used in ATD studies are likely to alter cognition. Meals can enhance memory, and that includes meals of carbohydrate, fat and protein Kaplan RJ, 2001. If protein meals can enhance cognition then amino acid mixtures may also have this effect. In addition there may be changes in cognition, from baseline to 5 hrs after amino acid administration, related to other factors such as the participants’ diurnal rhythms Blatter K, 2007, possible stress related to participation in the study, and so on. The issue that is not whether there will be changes in cognition in the control participants, but whether factors that may alter cognition, other than tryptophan availability, are different in the control and ATD conditions. The BCAA in the 100g ATD mixture and the relevant control mixture are the same, and it is therefore an appropriate control mixture.

The other main issue raised by Badawy and Dougherty concerns the lowering of the [Trp]/[LNAA] ratio in control mixtures. They mention a decline of 8% in this ratio with their recommended formulation and of around 40% with the control mixture for the 100g depletion mixture. On the other hand after ATD with the 100g mixture the decline in the [Trp]/[LNAA] ratio is well over 90% Predmore DB, 1976. As mentioned above, given that the [Trp]/[LNAA] ratio is a relatively poor index of tryptophan transport into brain, smaller changes in the ratio are not reliable. However, even if there were a decline in tryptophan availability for transport into brain of 40% would this alter serotonin function? Competition between the different amino acids for the transporter implies that the transporter is relatively close to saturation with amino acids. In this circumstance the actual change in transport of amino acids will be less than the change in plasma availability of the amino acids, as measured by plasma levels. Furthermore, tryptophan hydroxylase is about half saturated with tryptophan in human brain Young SN, 1981 so the decline in serotonin synthesis will be less than the decline in brain tryptophan. Will a relatively small decline in serotonin synthesis lead to a decrease in serotonin function? Little is known about this issue in human brain. However, for a decline in serotonin synthesis to alter serotonin function there must be a decline in serotonin release. This presumably implies a decrease in the serotonin content of vesicles. Another possibility is that a decline in synthesis leads to decreased catabolism of serotonin with no change in the vesicle serotonin content. Furthermore, given the presence of serotonin autoreceptors regulating, among other factors, the rate of firing of serotonin neurons, could a decrease in serotonin cause compensatory changes such an increase in neuronal firing? The factors mentioned above presumably explain why a very large decrease in plasma tryptophan availability is needed to see a lowering of mood Van der Does AJ, 2001, a change presumably mediated by lowered serotonin function. Thus, while the control mixture used with the original 100g ATD mixture probably does decrease the availability of tryptophan to the brain by a small extent, the decline is very much less than that needed to see a decline in mood. Furthermore, as pointed out in my first comment variations in tryptophan availability, as measured by the plasma ratio, vary over a 2-fold range under normal circumstances. Badawy and Dougherty provide no information demonstrating that the control mixture used with the original 100g ATD mixture causes any change, due to changes in serotonin function, in mood, behavioral or cognition. Badawy and Dougherty’s hypothesis could be tested by varying the tryptophan content of the control mixture to see if the original control mixture causes different outcomes in measures of mood, behavior or cognition, relative to a mixture that is identical except that it contains enough tryptophan to keep the [Trp]/[LNAA] ratio constant over time. I would be interested to see the results of such a study.

In relation to side effects, Badawy and Dougherty mention again one of their studies where there were side effects, and attrition of women, with the 100g ATD mixture. However this is only one study. They do not mention the fact that others had no attrition in studies on women with the 100g ATD mixture, e.g. Yatham LN, 2001. To say that one mixture is significantly better than another as far as side effects or attrition is concerned this would have to be demonstrated using a direct comparison of the different mixtures in the same study. I would welcome such a study comparing the original 100g ATD mixture, and Badawy and Dougherty’s glycomacropeptide based mixture.

In summary, I do not think the criticisms of Badawy and Dougherty are based on firm evidence.

On 2016 Jan 23, Abdulla A-B Badawy commented:

Abdulla A-B Badawy and Donald M Dougherty We acknowledge Professor Young’s important contribution to this field and our recommendations regarding the use of amino acid mixtures are based on his formulation. We proposed lowering the branched-chain amino acid (BCAA) content by 40% (from 30% to 18%) to normalise the [Trp]/[CAA] and the [Tyr + Phe]/[BCAA + Trp] ratios in the control formulation, with appropriate deletions or additions for depletion and loading, on the basis of a review Badawy AA, 2010 of many studies demonstrating Trp depletion by the control formulation of Young SN, 1985 and our experiments with various concentrations of BCAA Badawy AA, 2010. The decrease in the [Trp]/[CAA] ratio by the control formulation led investigators to increase its Trp content from the usual 2.3g (per 100g) to 2.9, 3.1, 4.1 or 4.6g (referenced in Badawy AA, 2015). However, the Trp ratio is still decreased with the 3.1g addition and increased by the larger additions. Young argues this point by quoting data obtained with proteins rather than amino acid mixtures, and although his formulation is based on the composition of human milk, it is still an amino acid formulation. The use of a low ATD formulation dose (25g) as a conservative control may reduce type 1 error, but its BCAA content can influence cognitive behaviour and thus minimise the effect of the larger BCAA dose. Reducing the BCAA content not only normalises the Trp and Tyr ratios Badawy AA, 2010, but can also minimise the BCAA effects on cognition and hence potential modulation of behavioural outcomes. Regarding our criteria for an ideal formulation, we agree with Young that his is the only ATD method validated by CSF measurements. By “robust”, we meant biochemical and not behavioural or other parameters. In the study by Moreno FA, 2010 however, CSF [Trp] did not correlate with serum Trp, but rather with the [Trp]/[LNAA] ratio. No data on serum Tyr or Phe were provided and the sham control treatment lowered the [Trp]/[LNAA] ratio by 25% at 5h. In our study Badawy AA, 2010 the corresponding decrease with the same formulation was 36% at 5h and slightly stronger (40-45%) at earlier time intervals. By contrast, only an 8% decrease in the ratio is observed with our recommended formulation. <PMID: 19896342 > calculated the ratio based on amino acid concentrations in µg/ml rather than µM. Because Trp has the largest molecular mass among LNAA, their calculation will have exaggerated [Trp] and minimised [LNAA], thereby minimising the decrease in the ratio. Our “tolerable” side effects criterion applies to somatic bodily symptoms rather than mood or cognition and drop-outs in our study involved only females receiving the 100g formulation for ATD and ATL (acute Trp loading) Dougherty DM, 2008 .Ours is therefore the only study demonstrating that a 50g dose of the ATD or ATL formulation does not cause attrition. This female sensitivity has previously been reported Sobczak S, 2014 and many studies have demonstrated the impact of somatic symptoms on attrition Dougherty DM, 2008. As Young suggests, 5-HT receptors and bright light may be factors and these and other potential mediators clearly require investigation. We also agree with Young that other factors including interactions between participants and with research staff can impact behavioural outcomes. Our approach is specific for the biochemical effects of the formulations and has to be considered along with other strategies.<br> Finally, we should like to address the question of specificity of the Young SN, 1985 formulation for serotonin. From the data in Table 3 by Leyton M, 2000, the [Trp]/[LNAA], [Phe]/[LNAA], [Tyr]/[LNAA] and [Phe + Tyr]/[BCAA + Trp] ratios are decreased by the control (balanced) formulation at 5h by 41%, 40%, 29% and 38% respectively. Many other studies have reported decreases in control formulations in the [Trp]/[CAA] ratio of up to 61% and in the [Phe + Tyr]/[BCAA + Trp] ratio of 40-60% (see Badawy AA, 2010). Thus, more studies have demonstrated decreases in the above ratios, compared to the 3 quoted by Young demonstrating no change. A recent ATD and ATPD (acute Tyr and Phe depletion) study Hildebrand P, 2015 in which the content of BCAA was 21.8% (closer to our recommended 18% than the traditional 30% of Young SN, 1985) showed that the control formulation did not alter the [Trp]/[CAA] ratio. Thus, until we establish why ratios are decreased in some, but not other, studies, it is prudent to aim for accuracy and ensure that ratios are not altered by a supposedly balanced control formulation. In subjects with normal but borderline ratios, a 30-40% decrease in ratios may tilt the balance to a state of depletion sufficiently to influence behavioural measures and their quantification in test subjects. In the study by Hildebrand P, 2015, the amino acid mixture was administered based on body wt: a further important aspect of standardization of the methodology. The control formulation for ATD and ATPD used by these latter authors included only 8 amino acids (instead of the 15 used by Young), which were the 3 BCAA plus Trp, Phe, Met, Threo and Lys. This control formulation with a Phe content of 9.24 g/70 kg, (compared with 5.7g in the <PMID: 3931142 > formulation) decreased the Tyr ratio by ~36% at 5h, thus illustrating the ability of Phe loading to decrease Tyr availability to the brain. This is one potential mechanism of the defective catecholamine synthesis in phenylketonuria (PKU). We do not doubt the low cerebral activity of Phe hydroxylase in PKU, although rates could be as high as 15-18% of those in controls van Spronsen FJ, 1998. We merely suggested an additional mechanism of decreased catecholamine synthesis, that of possible inhibition of cerebral Tyr hydroxylase (TH) by any excess Tyr that could be formed from Phe by TH. Human recombinant TH exhibits an equal Vmax for Tyr and Phe, but its Km for Phe is 8-fold higher than for Tyr Martínez A, 1993. Compared with controls, PKU patients exhibit a 6.4- and 4.6-fold higher plasma and CSF [Phe] respectively, a 1.8-fold lower plasma, but a 2.3-fold higher CSF, [Tyr] Ratzmann GW, 1984. Thus, whereas the plasma to CSF ratio of Phe in PKU resembles that of controls, the corresponding ratio for Tyr is significantly lower in PKU. The authors suggested that this may result from intracellular changes in brain metabolism and we suggest that Phe hydroxylation by TH may be responsible for this relative CSF Tyr elevation. Inhibition of catecholamine synthesis in PKU could therefore also result from substrate inhibition of TH by the likely excess of Tyr formed from Phe as suggested by us and indeed, as suggested by Young, by Phe inhibition of TH.

On 2016 Jan 10, Simon Young commented:

Badawy and Dougherty raise a number of important points in this paper, but there are number of issues that need discussion.

1. The authors suggest that the original 100g amino acid mixture used for acute tryptophan depletion (ATD), which remains the most commonly used mixture in ATD studies, lacks specificity and causes a lowering of dopamine and possibly norepinephrine. The authors say the 100g mixture lacks specificity because the mixture causes a decrease in the plasma [Phe + Tyr]/[CAA] ratio (an index of the transport of the catecholamine precursors into brain) thereby decreasing Tyr availability to the brain. However, they neglect to mention a study that investigated this issue through measurements on cerebrospinal fluid (CSF). Samples of CSF were taken before and after healthy volunteers received the 100g tryptophan-deficient amino acid mixture Carpenter LL, 1998. CSF tyrosine remained unchanged for the first few hours after the amino acid mixture, but rose slightly between about 5 and 10 hours after the mixture. CSF levels of the dopamine metabolite homovanillic acid (HVA) remained unchanged. Why do tyrosine and HVA not decline if the [Phe + Tyr]/[CAA] ratio decreases? Rat brain tyrosine hydroxylase can hydroxylate both phenylalanine and tyrosine, but it is not known whether this is true for the human brain enzyme. However, results from patients with phenylketonuria (PKU), who have very high phenylalanine levels, suggest that the activity of the human brain enzyme towards phenylalanine is not functionally significant. Thus, untreated PKU patients with phenylketonuria have low CSF tyrosine and HVA levels. When patients are treated with a low phenylalanine diet, or with tyrosine, CSF HVA levels increase Lykkelund C, 1988 Lou HC, 1985. Thus, the appropriate ratio to use as an index of tyrosine availability for catecholamine synthesis is [Tyr]/[CAA]. Badawy has stated that [Tyr]/[CAA] declines in a similar way to [Phe + Tyr]/[CAA] after ATD Badawy AA, 2013. However, others have reported that the control and tryptophan-depleted amino acid mixture cause the [Tyr]/[CAA] ratio to decrease slightly Leyton M, 2000, or not at all Golightly KL, 2001. These latter results are consistent with the CSF studies mentioned above. Although the conclusion is not definitive, the results available at this time suggest that the original 100g amino acid mixture does not decrease catecholamine synthesis and there is no convincing evidence it lacks specificity. Thus, there is no need to decrease the CAA in the 100g formula.

2. The authors suggest that the control mixture (the 100g ATD mixture plus 2.3g tryptophan) is inappropriate because it lowers the plasma [Trp/CAA] ratio. A study in which healthy participants were fed diets containing different levels of protein, and their blood was taken during the course of a day, demonstrated that the plasma [Trp/CAA] ratio varies over a two-fold range depending on the protein content of the meals and diurnal variation Fernstrom JD, 1979. The original ATD control mixture is based on the amino acid profile in human milk, so its effects would not be unphysiological to any important extent Young SN, 2013. There is a big separation between the effect of the control and depleting amino acid mixtures, as the level of CSF tryptophan after the depletion mixture is only 16.3% of the level after the control mixture Moreno FA, 2010. Some studies have used a 25g ATD mixture as a control for the 100g ATD mixture and no mood changes were seen after the control treatment e.g. Booij L, 2005. Thus, using a control mixture that causes a small decline in the [Trp/CAA] ratio does not seem to be a problem and, being a conservative control, may decrease the chance of a type I error.

3. The paper suggests various criteria for the ideal amino acid depletion and loading formulations. One of these is “Robust and reproducible changes in the study parameters”. Presumably the “study parameters” refer to biochemical changes rather than changes in mood, cognition or other outcomes. The main parameters the paper refers to are changes in plasma levels of amino acids, and in particular ratios such as [Trp/CAA]. The [Trp/CAA] is a rough index of the uptake of Trp into brain, which will be rough index of the brain tryptophan level, which is only one of several factors that regulates brain serotonin synthesis. A more valid method to obtain information is to look at measures related more directly to brain tryptophan levels and brain serotonin synthesis. An index of these can be obtained by the measurement of tryptophan and 5-HIAA in CSF. The original 100g ATD mixture has been shown to decrease human brain tryptophan and serotonin synthesis in 5 studies using measurements on CSF Carpenter LL, 1998, Moreno FA, 2000, Moreno FA, 2010, Salomon RM, 2003, Williams WA, 1999, and to decrease human brain serotonin synthesis in one study using a positron emission tomography method Nishizawa S, 1997. Currently no other method for ATD has been validated in this way. Another criterion is “Freedom from, or acceptable (tolerable), side effects not leading to attrition”. In discussing this issue the authors fail to mention two important issues. First, serotonin receptors in the gut and brain are involved in nausea and emesis Hasler WL, 1999. Second, bright light, relative to dim light, decreases side effects after ATD aan het Rot M, 2008. This suggests that central changes may be partly responsible for nausea, and that a greater lowering of serotonin, which may be more likely to reveal effects of ATD on mood, behavior or cognition, may also be more likely to induce nausea. Furthermore, there is currently no direct comparison of the 100g mixture with any alternative mixture, so direct evidence that the 100g mixture causes greater side effects is lacking.

4. Standardizing the amino acid formulation for ATD is only one factor that will help to increase the comparability of research reported in different studies. Some important factors have not always been controlled, while others cannot be controlled across studies. For example, the expectations that participants will have, that might influence their responses, could be influenced by the wording of the consent form and the demeanor and empathy of the person interacting with the participants. Other factors can be controlled. For example, bright light can reverse mood effects of ATD aan het Rot M, 2008 so studies should not be carried out in rooms with windows that can let in sunlight, and the light level should be standardized at a particular level. I suggest 200-300 lux. Between taking the amino acid mixture and being tested there is a period of several hours. If the participants are allowed to interact with others during this period this may influence their mood. For most people agreeable interactions are associated with better mood and quarrelsome interactions are associated with worse mood Côté S, 1998. Therefore interactions between research participants and research staff should be minimized. This can be done by keeping research participants by themselves in a room, and minimizing their interactions with research staff, while giving them access to relatively affectively neutral reading material and movies to avoid boredom between giving them the amino acid mixtures and testing. This has been done in my laboratory, e.g. Benkelfat C, 1994, and I suggest that this should also be standard practice in ATD studies.

In conclusion, the original ATD mixture is the only mixture that has been shown to lower serotonin synthesis in human brain, and criticisms of the method are not well supported. Currently it remains the best candidate for a standard mixture to use in ATD studies, although future studies may change this conclusion.

On 2016 Jan 10, Simon Young commented:

Badawy and Dougherty raise a number of important points in this paper, but there are number of issues that need discussion.

1. The authors suggest that the original 100g amino acid mixture used for acute tryptophan depletion (ATD), which remains the most commonly used mixture in ATD studies, lacks specificity and causes a lowering of dopamine and possibly norepinephrine. The authors say the 100g mixture lacks specificity because the mixture causes a decrease in the plasma [Phe + Tyr]/[CAA] ratio (an index of the transport of the catecholamine precursors into brain) thereby decreasing Tyr availability to the brain. However, they neglect to mention a study that investigated this issue through measurements on cerebrospinal fluid (CSF). Samples of CSF were taken before and after healthy volunteers received the 100g tryptophan-deficient amino acid mixture Carpenter LL, 1998. CSF tyrosine remained unchanged for the first few hours after the amino acid mixture, but rose slightly between about 5 and 10 hours after the mixture. CSF levels of the dopamine metabolite homovanillic acid (HVA) remained unchanged. Why do tyrosine and HVA not decline if the [Phe + Tyr]/[CAA] ratio decreases? Rat brain tyrosine hydroxylase can hydroxylate both phenylalanine and tyrosine, but it is not known whether this is true for the human brain enzyme. However, results from patients with phenylketonuria (PKU), who have very high phenylalanine levels, suggest that the activity of the human brain enzyme towards phenylalanine is not functionally significant. Thus, untreated PKU patients with phenylketonuria have low CSF tyrosine and HVA levels. When patients are treated with a low phenylalanine diet, or with tyrosine, CSF HVA levels increase Lykkelund C, 1988 Lou HC, 1985. Thus, the appropriate ratio to use as an index of tyrosine availability for catecholamine synthesis is [Tyr]/[CAA]. Badawy has stated that [Tyr]/[CAA] declines in a similar way to [Phe + Tyr]/[CAA] after ATD Badawy AA, 2013. However, others have reported that the control and tryptophan-depleted amino acid mixture cause the [Tyr]/[CAA] ratio to decrease slightly Leyton M, 2000, or not at all Golightly KL, 2001. These latter results are consistent with the CSF studies mentioned above. Although the conclusion is not definitive, the results available at this time suggest that the original 100g amino acid mixture does not decrease catecholamine synthesis and there is no convincing evidence it lacks specificity. Thus, there is no need to decrease the CAA in the 100g formula.

2. The authors suggest that the control mixture (the 100g ATD mixture plus 2.3g tryptophan) is inappropriate because it lowers the plasma [Trp/CAA] ratio. A study in which healthy participants were fed diets containing different levels of protein, and their blood was taken during the course of a day, demonstrated that the plasma [Trp/CAA] ratio varies over a two-fold range depending on the protein content of the meals and diurnal variation Fernstrom JD, 1979. The original ATD control mixture is based on the amino acid profile in human milk, so its effects would not be unphysiological to any important extent Young SN, 2013. There is a big separation between the effect of the control and depleting amino acid mixtures, as the level of CSF tryptophan after the depletion mixture is only 16.3% of the level after the control mixture Moreno FA, 2010. Some studies have used a 25g ATD mixture as a control for the 100g ATD mixture and no mood changes were seen after the control treatment e.g. Booij L, 2005. Thus, using a control mixture that causes a small decline in the [Trp/CAA] ratio does not seem to be a problem and, being a conservative control, may decrease the chance of a type I error.

3. The paper suggests various criteria for the ideal amino acid depletion and loading formulations. One of these is “Robust and reproducible changes in the study parameters”. Presumably the “study parameters” refer to biochemical changes rather than changes in mood, cognition or other outcomes. The main parameters the paper refers to are changes in plasma levels of amino acids, and in particular ratios such as [Trp/CAA]. The [Trp/CAA] is a rough index of the uptake of Trp into brain, which will be rough index of the brain tryptophan level, which is only one of several factors that regulates brain serotonin synthesis. A more valid method to obtain information is to look at measures related more directly to brain tryptophan levels and brain serotonin synthesis. An index of these can be obtained by the measurement of tryptophan and 5-HIAA in CSF. The original 100g ATD mixture has been shown to decrease human brain tryptophan and serotonin synthesis in 5 studies using measurements on CSF Carpenter LL, 1998, Moreno FA, 2000, Moreno FA, 2010, Salomon RM, 2003, Williams WA, 1999, and to decrease human brain serotonin synthesis in one study using a positron emission tomography method Nishizawa S, 1997. Currently no other method for ATD has been validated in this way. Another criterion is “Freedom from, or acceptable (tolerable), side effects not leading to attrition”. In discussing this issue the authors fail to mention two important issues. First, serotonin receptors in the gut and brain are involved in nausea and emesis Hasler WL, 1999. Second, bright light, relative to dim light, decreases side effects after ATD aan het Rot M, 2008. This suggests that central changes may be partly responsible for nausea, and that a greater lowering of serotonin, which may be more likely to reveal effects of ATD on mood, behavior or cognition, may also be more likely to induce nausea. Furthermore, there is currently no direct comparison of the 100g mixture with any alternative mixture, so direct evidence that the 100g mixture causes greater side effects is lacking.

4. Standardizing the amino acid formulation for ATD is only one factor that will help to increase the comparability of research reported in different studies. Some important factors have not always been controlled, while others cannot be controlled across studies. For example, the expectations that participants will have, that might influence their responses, could be influenced by the wording of the consent form and the demeanor and empathy of the person interacting with the participants. Other factors can be controlled. For example, bright light can reverse mood effects of ATD aan het Rot M, 2008 so studies should not be carried out in rooms with windows that can let in sunlight, and the light level should be standardized at a particular level. I suggest 200-300 lux. Between taking the amino acid mixture and being tested there is a period of several hours. If the participants are allowed to interact with others during this period this may influence their mood. For most people agreeable interactions are associated with better mood and quarrelsome interactions are associated with worse mood Côté S, 1998. Therefore interactions between research participants and research staff should be minimized. This can be done by keeping research participants by themselves in a room, and minimizing their interactions with research staff, while giving them access to relatively affectively neutral reading material and movies to avoid boredom between giving them the amino acid mixtures and testing. This has been done in my laboratory, e.g. Benkelfat C, 1994, and I suggest that this should also be standard practice in ATD studies.

In conclusion, the original ATD mixture is the only mixture that has been shown to lower serotonin synthesis in human brain, and criticisms of the method are not well supported. Currently it remains the best candidate for a standard mixture to use in ATD studies, although future studies may change this conclusion.

On 2016 Feb 18, Abdulla A-B Badawy commented:

None