JUDGMENT OF THE GENERAL COURT (Second Chamber)

18 October 2023 (*)

(Environment – Genetically modified products – Genetically modified soybean containing, consisting of or produced from MON 87751 × MON 87701 × MON 87708 × MON 89788 and its subcombinations – Rejection of an application for internal review of an authorisation for placing on the market – Manifest error of assessment – Article 10 of Regulation (EC) No 1367/2006 – Articles 4 to 6 and 16 to 18 of Regulation (EC) No 1829/2003 – Article 5 of Implementing Regulation (EU) No 503/2013)

In Case T‑606/21,

TestBioTech eV, established in Munich (Germany), represented by K. Smith, Barrister,

applicant,

v

European Commission, represented by I. Galindo Martín and F. Castilla Contreras, acting as Agents,

defendant,

THE GENERAL COURT (Second Chamber),

composed of A. Marcoulli, President, V. Tomljenović (Rapporteur) and W. Valasidis, Judges,

Registrar: V. Di Bucci,

having regard to the written part of the procedure,

having regard to the fact that no request for a hearing was submitted by the parties within three weeks after service of notification of the close of the written part of the procedure, and having decided to rule on the action without an oral part of the procedure, pursuant to Article 106(3) of the Rules of Procedure of the General Court,

gives the following

Judgment

1        By its action under Article 263 TFEU, the applicant, TestBioTech eV, seeks annulment of the decision of the European Commission of 8 July 2021 (‘the contested decision’) rejecting the grounds set out in its request for internal review made pursuant to Article 10(1) of Regulation (EC) No 1367/2006 of the European Parliament and of the Council of 6 September 2006 on the application of the provisions of the Aarhus Convention on Access to Information, Public Participation in Decision-making and Access to Justice in Environmental Matters to Community institutions and bodies (OJ 2006 L 264, p. 13), against Commission Implementing Decision (EU) 2021/66 of 22 January 2021 authorising the placing on the market of products containing, consisting of or produced from genetically modified soybean MON 87751 × MON 87701 × MON 87708 × MON 89788, pursuant to Regulation (EC) No 1829/2003 of the European Parliament and of the Council (OJ 2021 L 26, p. 44) (‘the authorisation decision’).

 Background to the dispute

2        On 17 December 2015, Monsanto Europe SA/NV, acting on behalf of Monsanto Company, United States (‘Monsanto’), submitted to the competent authority of the Netherlands, in accordance with Articles 5 and 17 of Regulation (EC) No 1829/2003 of the European Parliament and of the Council of 22 September 2003 on genetically modified food and feed (OJ 2003 L 268, p. 1), an application for the placing on the market of products containing, consisting of, or produced from genetically modified soybean MON 87751 × MON 87701 × MON 87708 × MON 89788 (‘the modified soybean’), intended for import, transformation and all uses like any other soybean, excluding cultivation, in the European Union.

3        The modified soybean is a genetically modified organism within the meaning of Article 2(5) of Regulation No 1829/2003, read in conjunction with Article 2(2) of Directive 2001/18/EC of the European Parliament and of the Council of 12 March 2001 on the deliberate release into the environment of genetically modified organisms and repealing Council Directive 90/220/EEC (OJ 2001 L 106, p. 1). It is a product obtained by conventional crossing which combines the genetic material of certain genetically modified plants containing single transformation events and which express the following proteins:

–        MON 89788 expressing the 5-enolpyruvylshikimate-3-phosphate synthase (CP4 EPSPS) protein; a plant that expresses the CP4 EPSPS protein has a low affinity for glyphosate, allowing such a plant to continue to grow when treated with glyphosate-based herbicides; plants without the CP4 EPSPS protein (for example, weeds found in agriculture) die when they are exposed to such herbicides;

–        MON 87708 expressing the DMO protein for tolerance to dicamba-containing herbicides; dicamba is the active ingredient in some agricultural herbicides; that herbicide kills plants by mimicking natural plant hormones known as auxins, which regulate many plant processes including cell growth and protein synthesis; it can induce rapid cell growth and limit transpiration and photosynthesis in plants, leading to irregular plant growth, leaf drop and starvation; plants expressing the DMO protein are capable of degrading dicamba and so are tolerant of it;

–        MON 87751 expressing proteins Cry1A.105 and Cry2Ab2; the proteins known as ‘Cry’ are toxins derived from the bacterium Bacillus thurigiensis, a soil-dwelling bacterium; commonly known as ‘Bt toxins’, Cry toxins confer protection against certain lepidopteran pests;

–        MON 87701 expressing the insecticidal protein Cry1Ac.

4        The modified soybean thus combines the insecticidal traits of MON 87751 and MON 87701 with the herbicide-tolerant traits of MON 89788 and MON 87708. As it combines the transgenes of parents which are single transformation events, the modified soybean is what is known as a ‘stacked event’.

5        On 25 September 2019, the Scientific Panel on Genetically Modified Organisms (GMO) of the European Food Safety Authority (EFSA) (‘the GMO Scientific Panel’) issued an opinion on the application for authorisation (‘EFSA’s opinion of 25 September 2019’), the conclusion of which was that – so far as concerns its potential effects on human and animal health and the environment – the modified soybean was as safe as and nutritionally equivalent to its non-genetically modified comparator and the commercial non-genetically modified reference varieties tested.

6        On 22 January 2021, as is apparent from Article 2 of the authorisation decision, the Commission authorised, under certain conditions, for the purposes of Article 4(2) and Article 16(2) of Regulation No 1829/2003:

(a)      foods and food ingredients containing, consisting of, or produced from the modified soybean;

(b)      feed containing, consisting of or produced from the modified soybean;

(c)      products containing or consisting of the genetically modified soybean for uses other than those provided for in subparagraphs (a) and (b), with the exception of cultivation.

7        By a letter dated 27 August 2018, Monsanto Europe informed the Commission that it had changed its legal form and its company name to Bayer Agriculture BVBA, Belgium. By letter of 28 July 2020, it informed the Commission that, as from 1 August 2020, it would change its legal form and its company name to Bayer Agriculture BV, Belgium. By letter of 28 July 2020, it informed the Commission that, as from 1 August 2020, Monsanto would also change its legal form and its company name to Bayer CropScience LP, United States.

8        On 8 March 2021, the applicant, which is a not-for-profit association seeking to promote research in the field of biotechnology, submitted a request for internal review of the authorisation decision under Article 10(1) of Regulation No 1367/2006 (‘the request for internal review’). The Commission requested EFSA to provide technical assistance, by conducting an in-depth analysis of the scientific aspects of that request.

9        On 3 May 2021, EFSA published a technical report on the detailed scientific grounds provided by the applicant in the request for internal review (‘the technical report’). In that report, entitled ‘Scientific assistance on the internal review under Regulation (EC) No 1367/2006 of the Commission Implementing Decisions on genetically modified soybean MON 87751 x MON 87701 x MON 87708 x MON 89788, maize MON 87427 x MON 87460 x MON 89034 x MIR162 x NK603 and subcombinations and maize MON 87427 x MON 89034 x MIR162 x MON 87411 and subcombinations’, EFSA concluded that that application did not justify reassessing its opinion of 25 September 2019 and its risk management recommendations concerning the modified soybean.

10      On 8 July 2021, the Commission replied to the request for internal review by the contested decision. In that decision, to which was attached, as Annex I, the Commission’s detailed assessment of the grounds for review relied on (‘the detailed assessment’), the Commission considered, in essence, that the grounds relied on by the applicant in its request for internal review, including the detailed technical background provided, did not justify a re-examination of the conclusions and risk management recommendations of the GMO Scientific Panel on the modified soybean.

 Forms of order sought

11      The applicant claims that the Court should:

–        annul the contested decision;

–        order the Commission to pay the costs.

12      The Commission contends that the Court should:

–        dismiss the action;

–        order the applicant to pay the costs.

 Law

13      In support of the present action, the applicant raises a single plea in law. Before providing an answer to that plea, it is necessary to set out certain elements of the relevant legal framework.

 Preliminary observations

14      It is apparent from Article 10(1) of Regulation No 1367/2006, in the version in force on the date on which the request for internal review was made, namely 8 March 2021, that any non-governmental organisation which meets the criteria set out in Article 11 of that regulation is entitled to make a request for internal review of administrative acts of individual scope, having legally binding and external effects, and that request must state the grounds for the review.

15      In order to set out the grounds for conducting the review in the manner required, a party requesting the internal review of an administrative act under environmental law is required to put forward the facts or legal arguments of sufficient substance to give rise to substantial or serious doubts as to the assessment made in that act by the EU institution or body (see, to that effect, judgments of 12 September 2019, TestBioTech and Others v Commission, C‑82/17 P, EU:C:2019:719, paragraph 69; of 15 December 2016, TestBioTech and Others v Commission, T‑177/13, not published, EU:T:2016:736, paragraphs 67, 83 and 88; and of 4 April 2019, ClientEarth v Commission, T‑108/17, EU:T:2019:215, paragraph 57).

16      Where the Commission concludes that the evidence adduced by a party requesting an internal review is substantial and liable to raise serious doubts as to the formal or substantive lawfulness of the grant of that authorisation, it is required to examine all relevant information of its own motion (judgments of 15 December 2016, TestBioTech and Others v Commission, T‑177/13, not published, EU:T:2016:736, paragraph 85, and of 4 April 2019, ClientEarth v Commission, T‑108/17, EU:T:2019:215, paragraph 216). At the end of its review, the Commission may withdraw or amend the authorisation decision, thereby granting in whole or in part the request for review, or maintain the authorisation as it was granted, by rejecting the request.

17      If the request for internal review is rejected, as in the present case, the party which made that request may, in accordance with Article 12 of Regulation No 1367/2006, read in conjunction with Article 10 thereof, bring an action – on grounds of lack of competence, infringement of an essential procedural requirement, infringement of the Treaties or of any rule of law relating to their application, or misuse of powers – against the decision rejecting that request as unfounded before the Courts of the European Union in accordance with Article 263 TFEU (judgment of 12 September 2019, TestBioTech and Others v Commission, C‑82/17 P, EU:C:2019:719, paragraph 38).

18      The scope of the judicial review carried out by the Courts of the European Union in their examination of an action for annulment brought pursuant to Article 12 of Regulation No 1367/2006 against a decision of the Commission concerning a request for internal review adopted pursuant to Article 10 of that regulation does not differ from the scope of the judicial review which the General Court carries out on the merits of the grounds of decisions directly challenged on the basis of Article 263(2) and (4) TFEU (see, to that effect, judgment of 15 December 2016, TestBioTech and Others v Commission, T‑177/13, not published, EU:T:2016:736, paragraph 81).

19      In that regard, it must be borne in mind that, where an EU institution is called upon to make complex assessments, it enjoys a wide measure of discretion, the exercise of which is subject to a judicial review restricted to verifying that the measure in question is not vitiated by a manifest error or a misuse of powers and that the competent authority did not clearly exceed the bounds of its discretion (see judgment of 15 December 2016, TestBioTech and Others v Commission, T‑177/13, not published, EU:T:2016:736, paragraph 77 and the case-law cited).

20      In order to establish that an institution committed a manifest error in assessing complex facts such as to justify the annulment of an act, the evidence adduced by the applicant must be sufficient to make the factual assessments used in the act implausible. Subject to that review of plausibility, it is not the Court’s role to substitute its assessment of complex facts for that made by the institution which adopted the decision. Consequently, a plea alleging the existence of a manifest error must be rejected if, despite the evidence adduced by the applicant, the contested assessment may still be accepted as true or valid. That is particularly so where the decision at issue is vitiated by errors which, taken together, are of only minor significance unlikely to have influenced the administration (see judgment of 4 April 2019, ClientEarth v Commission, T‑108/17, EU:T:2019:215, paragraph 246 and the case-law cited).

21      Furthermore, the limits to the review by the Courts of the European Union do not affect their duty to establish whether the evidence relied on is factually accurate, reliable and consistent, whether that evidence contains all the relevant information which must be taken into account in order to assess a complex situation, and whether it is capable of substantiating the conclusions drawn from it (see judgment of 15 December 2016, TestBioTech and Others v Commission, T‑177/13, not published, EU:T:2016:736, paragraph 79 and the case-law cited).

22      Lastly, where the EU institutions have a broad discretion, respect for the rights guaranteed by the EU legal order in administrative procedures is of even more fundamental importance. Those guarantees include, in particular, the duty of the competent institution to examine carefully and impartially all the relevant aspects of the individual case, the right of the person concerned to make his or her views known and also his or her right to have an adequately reasoned decision (see judgment of 15 December 2016, TestBioTech and Others v Commission, T‑177/13, not published, EU:T:2016:736, paragraph 80 and the case-law cited).

 The single plea in law

23      In the application, the applicant claims that there are ‘manifest errors of assessment’ stemming from the fact that the authorisation decision had been ‘confirmed’ without ensuring that Monsanto had provided, in the context of the procedure which led to the adoption of that decision, the appropriate data under Article 5(3)(f), Article 6(3)(a), Article 17(3)(f) and Article 18(3)(a) of Regulation No 1829/2003 and Article 5 of Commission Implementing Regulation (EU) No 503/2013 of 3 April 2013 on applications for authorisation of genetically modified food and feed in accordance with Regulation No 1829/2003 and amending Commission Regulations (EC) No 641/2004 and (EC) No 1981/2006 (OJ 2013 L 157, p. 1). In addition, the Commission failed to ensure that EFSA had carried out an adequate risk assessment of the ‘highest possible standard’, focusing in particular on the question whether the stacking of genes presented by the modified soybean has potential effects on the gene expression of that plant when it is exposed to herbicides. Similarly, the Commission failed to give any adequate consideration to the potential for toxicity and allergenicity of the modified soybean because of the synergistic effects between the proteins whose expression results from the genetic modification, naturally occurring protease inhibitors in soybean and exposure to herbicides or to herbicide residues in harvests. Finally, the Commission failed to request that animal feeding tests of the modified soybean be conducted.

24      Worded in that way, the vast majority of those arguments are ineffective. The question which may arise in the context of an action such as that at issue is not the question whether the Commission was required to ‘ensure’ that the applicant for authorisation or EFSA complied with their obligations deriving from Regulation No 1829/2003 and Implementing Regulation No 503/2013 in the context of the procedure which led to the adoption of the authorisation decision. Similarly, the question cannot be whether the Commission failed to request that certain studies be carried out. By contrast, in a case such as that at issue, the question whether and how the Commission responded to the matters raised in the request for internal review is relevant (see paragraph 15 above).

25      In response to a measure of organisation of procedure, the applicant stated that, by its single plea in law, it was in fact alleging infringement of the substantive requirements laid down in Article 4(1)(a) and Article 16(1)(a) of Regulation No 1367/2006 – which form part of all the conditions which an authorisation must satisfy in accordance with Article 4(2), (3) and (5) and Article 16(2), (3) and (5) of Regulation No 1367/2006, and of all the conditions in the light of which the examination enabling the adoption of a decision such as the contested decision must be carried out.

26      An interpretation of the claims raised in the application by reference to the right to an effective remedy, guaranteed by Article 47(1) of the Charter of Fundamental Rights of the European Union, requires that they be regarded as having been put forward in support of a single plea in law alleging infringement of Article 4(1)(a) and Article 16(1)(a) of Regulation No 1367/2006, read in conjunction with Article 5(a) of Implementing Regulation No 503/2013 and with certain provisions contained in Annex II to that regulation, as referred to in the application.

27      In the present case, the requirement in Article 4(1)(a) and Article 16(1)(a) of Regulation No 1367/2006 that food and feed must not have adverse effects on human health, animal health or the environment is the fundamental material assessment criterion on the basis of which an authorisation such as that at issue must be assessed, in accordance with Article 4(2), (3) and (5) and Article 16(2), (3) and (5) of Regulation No 1367/2006; the same is true of a decision based on Article 10(2) of Regulation No 1367/2006, such as the contested decision.

28      The single plea in law concerns two distinct factual contexts, namely, first, exposure to the application of herbicides and, second, the assessment of allergenicity and toxicity, which it is appropriate to address as two different parts of that single plea.

 The first part, concerning exposure to the application of herbicides

29      The applicant submits, in essence, that the Commission’s answers in the contested decision to issues raised in the request for internal review regarding the possibility that drought may have an adverse effect on the gene expression of the modified soybean are vitiated by four manifest errors of assessment, the existence of which equates to an infringement of Article 4(1)(a) and Article 16(1)(a) of Regulation No 1367/2006, read in conjunction with Article 5(a) of Implementing Regulation No 503/2013 and with Sections 1.2.2.3, 1.3.1 and 1.3.2.1(b) of Part II of Annex II to Implementing Regulation No 503/2013.

30      In the first place, the Commission’s response, according to which, in the context of the field trials for the comparative analysis of genetically modified herbicide-tolerant plants, the intended herbicide doses should have been maintained ‘at a similar application rate across sites’ to ensure their comparability, is incorrect. That response shows that the field trials conducted in the present case did not meet the requirements of Implementing Regulation No 503/2013. If the application of the complementary intended herbicide in the fields of genetically engineered plants were to be maintained at a similar rate as that applied ‘in the fields with non-genetically engineered plants’, the conditions under which field testing was conducted would in no way represent actual growing conditions. According to the applicant, under real agricultural practices, significantly higher rates of the complementary intended herbicides are sprayed on herbicide-tolerant genetically engineered plants in comparison to non-genetically engineered plants which are not made tolerant to those herbicides. In the present case, the premiss should have been that, in practice, plants are exposed to high and repeated doses of glyphosate on its own or in conjunction with dicamba. When the modified soybean is exposed to higher doses of herbicide, the plants can experience stress conditions, impacting gene expression and plant composition as well as the biological characteristics of the modified soybean.

31      In the second place, given that the dose of glyphosate applied in the present case was only 0.87 kilograms (kg) (acid equivalent) per hectare, the Commission was wrong to agree, in Section 1.2.2 of the detailed assessment, with EFSA’s conclusion that the dose applied for the intended herbicide was consistent with the manufacturers’ recommendations. In the present case, the dose of intended herbicide applied in the Monsanto field trials was lower than the dose recommended by the manufacturer. First, on its own product label ‘Roundup Ready Soybean’, Monsanto recommends spraying a maximum application rate of 5.3 quarts per acre, which is equivalent to approximately 8.2 kg (active substance) per hectare. The total in-crop recommended application maximum is 64 fluid ounces per acre per year, which is equivalent to approximatively 3.1 kg (active substance) of glyphosate per hectare.

32      In any event, under Section 1.3.2.1(b) of Part II of Annex II to Implementing Regulation No 503/2013, field trials should be conducted taking into account real agricultural practices and not simply in accordance with a manufacturer’s recommendations. The dose of intended herbicide applied in the Monsanto field trials was substantially lower than would be expected under real conditions. A dose of 0.87 kg (acid equivalent) per hectare differs significantly from manufacturers’ recommendations. Second, in actual growing conditions in the United States, an average dose of 6 to 7 kg in total per hectare is expected to be applied, as is apparent in particular from the USDA (2019) and Miyazaki et al. (2019) studies. The application of even higher doses can be expected in South America, as is apparent from certain studies, such as the Miyazaki et al. (2019) and Avila-Vazquez et al. (2018) studies.

33      Similarly, the expression data from the modified soybean were generated from crops treated with the ‘complementary’ intended herbicide. Next, they were compared to data from the parental plants which were not treated with the ‘complementary’ intended herbicide. That approach amounts to a failure to comply with the requirements set out in Section 1.3.1 of Part II of Annex II to Implementing Regulation No 503/2013, according to which herbicide-resistant plants must be exposed to the ‘intended herbicide’ and compared on that basis.

34      Finally, the Commission’s interpretation of Implementing Regulation No 503/2013 would permit field trials of the relevant herbicide at minute levels (for example at 1% of the dosage to be applied in practice), which is however absurd.

35      In the third place, the applicant submits that it is true that the Commission accepted that the scientific literature cited by the applicant in the request for internal review, in particular the Zanatta et al. (2020) study, was relevant to the question of whether gene expression would be affected by high doses of herbicides, however, it concluded nonetheless that the gene expression data generated in the field testing were adequate. In that regard, the Commission referred to Section 3.5.6 of EFSA’s opinion of 25 September 2019, without, however, providing any reasoning of its own. In Section 3.5.6 of its opinion, EFSA set out the compositional data generated in flawed field testing. That does not in any way justify the failure to assess properly the impact of higher doses of herbicide.

36      In the fourth place, the Commission set aside certain publications cited by the applicant in its request for internal review by taking the view that they were not relevant because the data presented therein had been generated using material produced in a glyphosate-free environment, or using material not produced from a specific event, or because the data presented were generated from materials not produced from seeds or grains. However, the publications cited by the applicant establish (i) that herbicide application affects plant composition and metabolism in herbicide-resistant plants and (ii) that there is a clear correlation between the application of different doses (including higher doses) of glyphosate and plant agronomic performance and composition.

37      The Commission contests the applicant’s arguments.

38      As a preliminary point, it should be noted that the applicant failed to identify the specific sections of the contested decision or of the detailed assessment which, in its view, are vitiated by errors. It must be considered, however, that the applicant refers to Section 1.2.2 of the detailed assessment.

39      By its arguments, the applicant did not, however, deprive the assessments set out in Section 1.2.2 of the detailed assessment of their plausibility. It follows that the applicant has not demonstrated the existence of manifest errors vitiating the contested decision and, therefore, there is no infringement of Article 4(1)(a) and Article 16(1)(a) of Regulation No 1367/2006, read in conjunction with Article 5(a) of Implementing Regulation No 503/2013 and with Sections 1.2.2.3 and 1.3.1. and Section 1.3.2.1(b) of Part II of Annex II thereto. This can be inferred from the following assessments.

40      In the first place, by the arguments referred to in paragraph 30 above, the applicant seeks a declaration, first, that the field trials carried out by Monsanto did not reflect the ‘expected’ or ‘actual’ growing conditions of the modified soybean, since they did not reflect the quantities of intended herbicide which are used for the modified soybean under the agronomic practice of the countries in which that plant will be grown (United States, certain countries in South America, and so forth). Furthermore, according to the applicant, the Commission was wrong to base its assessment on data from field trials carried out by applying a dose of intended herbicide in modified soybean fields which was similar to or of the same level as the dose of conventional herbicide applied in the fields of non-genetically modified plants. Those arguments must be rejected as unfounded.

41      First, the applicant’s premiss that the modified soybean may, in practice, be exposed to high doses of glyphosate and dicamba, since parental plants MON 89788 and MON 87708 express proteins for tolerance to herbicides containing glyphosate and dicamba, respectively, has not been established. Thus, the applicant has not put forward any evidence showing clearly what might be the expected conditions under which the modified soybean is to be grown, for the purposes of Section 1.2.2.3(e) and Section 1.3.2.1(b) of Part II of Annex II to Implementing Regulation No 503/2013, or the ‘expected agricultural practices’ for the modified soybean, within the meaning of the fourth paragraph of Section 1.3.1 of Part II of Annex II to Implementing Regulation No 503/2013. In particular, it has not been demonstrated that the application of glyphosate and dicamba at high doses corresponds in practice to the expected growing conditions for the modified soybean, which exist in countries where exporters of that plant to the European Union are located. Nor has it been demonstrated that such a practice will be necessary for farmers to benefit from the herbicide-resistant properties of the modified soybean, and that modified soybean growers cannot benefit from the herbicide resistance of the modified soybean if they use doses of the intended herbicide falling within the range recommended by the manufacturer. Moreover, the applicant does not claim that the use of the doses of glyphosate recommended by the manufacturer fall outside good agricultural practices. Nor does the applicant show that growers need to or will have to spray bigger volumes of the intended herbicide or spray it at increased frequencies.

42      In so far as – in other parts of the application – the applicant refers to certain studies (namely, the USDA (2019), Miyazaki et al. (2019) and Avila-Vazquez et al. (2018) studies), it must be held that, as regards the expected growing conditions of the modified soybean, those studies have no probative value for the purposes of the present case. Not only was it not the purpose of those documents to establish the expected growing conditions of the modified soybean in the United States or South America, but they also do not include any verifiable analysis concerning that issue. It is true that those studies are based on certain hypotheses so far as concerns certain growing conditions for the modified soybean in certain countries, but they do not demonstrate that those hypotheses are correct.

43      Second, in the field trials carried out by Monsanto, glyphosate was admittedly applied only to sites containing the modified soybean and not to its conventional counterpart. However, that approach is consistent with the fourth paragraph of Section 1.3.1 of Part II of Annex II to Implementing Regulation No 503/2013. It is apparent from that provision that the intended herbicide is applied only to the genetically modified plant and not to its conventional counterpart. The fourth paragraph of Section 1.3.1 of Part II of Annex II to Implementing Regulation No 503/2013 does not require that the doses of intended herbicide sprayed on genetically modified herbicide-tolerant plants be applied in doses higher than those of conventional herbicide sprayed on non-genetically modified plants. It cannot therefore be inferred from Section 1.3.1 of Part II of Annex II to Implementing Regulation No 503/2013 that it was wrong for the Commission to base the assessment on data obtained from field trials carried out by applying an intended herbicide in modified soybean fields that was similar to or of the same level as that applied in non-genetically modified plant fields. Nor does that method of applying the intended herbicide in question, namely glyphosate, conflict with Sections 1.2.2.3 and 1.3.2.1(b) of Part II of Annex II to Implementing Regulation No 503/2013.

44      In the second place, the arguments by which the applicant seeks a declaration that the field trials carried out by Monsanto were insufficient because the quantities of intended herbicide applied did not correspond to the manufacturer’s recommendations (see paragraphs 31 to 33 above) are not convincing.

45      It is true that Section 1.2.2.3(e) and Section 1.3.2.1(b) of Part II of Annex II to Implementing Regulation No 503/2013 refer, in essence, to the need for the genetically modified plant to be examined in field trials carried out taking account of the expected conditions under which the modified soybean is to be grown. However, those provisions do not contain any requirement as regards the quantities of the intended herbicide – such as glyphosate and dicamba – which must be applied specifically to genetically modified plants.

46      The applicant has not provided any evidence that the quantities of intended herbicide applied in the present case did not correspond to the growing conditions of the modified soybean. The applicant’s reference to the label of the product ‘Roundup Ready Soybean’ is of no assistance in that context. It is true that that product label contains information and recommendations from the producer concerning the application of glyphosate for certain agricultural crops. However, it cannot be assumed that all the recommendations in that document actually reflect the expected growing conditions of the modified soybean. That document shows, at most, the quantities recommended by the glyphosate producer for genetically modified plants, listed on that product label. The fact that the product label has no probative value as regards the expected growing conditions for the modified soybean is clear from its content. The recommendations it contains are accompanied by a disclaimer, which provides, inter alia, that the ‘state or locality may require additional precautions and instructions for use of this product that are not included here’, ‘the information found [on] this label may differ from the information found on the product label’ and users are required to ‘follow the precautions and instructions for use on the label of the pesticide [they] are using’. It follows that the expected growing conditions of the modified soybean are dependent on certain additional safety instructions and certain instructions which are described on the product labels directly attached to glyphosate containers. Those instructions are additional to industry recommendations and form part of the expected growing conditions. The applicant does not state what those additional instructions might be. In the absence of precise information on those additional instructions and, therefore, ultimately, on the growing conditions of the modified soybean, any attempt to compare the level of intended herbicide applied in the field trials carried out in the present case with the dose of intended herbicide which could apply under expected growing conditions is impossible. It cannot therefore be found that the quantities of glyphosate applied in the field trials were considerably lower than those foreseen to apply in practice.

47      Furthermore, the applicant has also not shown that the quantities of intended herbicide applied in the Monsanto field trials were lower than the dose recommended by the manufacturer for the use of the intended herbicides in question, namely glyphosate and dicamba. As regards glyphosate, the applicant’s interpretation of the recommendations on the label of the product ‘Roundup Ready Soybean’, so far as concerns the expected growing conditions of the modified soybean, cannot be accepted. The value of 5.3 quarts per acre, which is equivalent to approximately 8.2 kg (active substance) per hectare, referred to by the applicant, and which corresponds to 6.7 kg (acid equivalent) per hectare, concerns a combined application on the soybean over a whole year and includes the amount of herbicide applied during treatments at various stages of cultivation, namely, pre-plant, in-crop and pre-harvest. However, it follows from the label of the product ‘Roundup Ready Soybean’ that the recommended dose, at a certain growth stage of the soybean, falls between 22 and 44 fluid ounces per acre, which is equivalent to a range falling between 0.86 and 1.7 kg (acid equivalent) per hectare. It follows that the quantity of 0.87 kg (acid equivalent) per hectare, applied ‘in-crop’ in the context of field trials, for the modified soybean is within the range recommended by the manufacturer. Therefore, the Commission cannot be held to have erred so far as concerns the application of Section 1.2.2.3(e) and Section 1.3.2.1(b) of Part II of Annex II to Implementing Regulation No 503/2013.

48      Nor has the applicant’s claim that – under the growing conditions of the modified soybean specific to the United States or South America – an average dose of 6 to 7 kg in total is to be applied per hectare been demonstrated. It is true that the USDA (2019), Miyazaki et al. (2019) and Avila-Vazquez et al. (2018) studies referred to by the applicant in support of that assertion were carried out on the basis of certain quantities of glyphosate applied at certain stages of growth of certain plants, but the quantities referred to in those studies are a hypothesis of scientific work. Thus, those studies, for their part, do not amount to evidence relating to the question of which quantities will actually be applied in the agronomic practice existing in the United States or in South America. As has been noted in paragraph 42 above, as regards the amounts of intended herbicide used in ‘actual’ or ‘expected’ growing conditions of the modified soybean, those studies simply have no probative value for the purposes of the present case.

49      Furthermore, as regards the applicant’s complaint that data on the expression of the modified soybean generated from plants treated with the intended herbicide were simply compared with data generated from parental plants which had not been treated with the intended herbicide, it cannot be held that Section 1.3.1 of Part II of Annex II to Implementing Regulation No 503/2013 (see paragraph 33 above) was infringed.

50      First, the applicant appears to derive its argument concerning the comparison between plants treated with the intended herbicide with the parental plants which were not treated with the intended herbicide from a document which Monsanto had annexed to its application for authorisation. That document, which is entitled ‘Application for authorisation, Part II “Scientific Information” EFSA-GMO-NL-2016-128’, forms part of the Court’s file as Annex 2 to the application. However, in the absence of an express and precise indication in the body of the application itself, the Court is not in a position to ascertain in which part of that document – which, moreover, contains approximately 200 pages – the comparison criticised by the applicant can be found. It is therefore not possible to understand how the data referred to by the applicant had been compared or, ultimately, the impact that that comparison may have had on the assessment the Commission was to carry out regarding the characteristics of the modified soybean. Consequently, the assertion referred to in paragraph 33 above must be rejected as being in no way substantiated.

51      Second, it is true that Section 1.3.1 of Part II of Annex II to Implementing Regulation No 503/2013 sets out certain procedures which must be followed for performing the comparative analysis. However, that provision does not preclude data from procedures and field trials, other than those set out in that provision, from being taken into account in the comparative analysis. Even if it is not expressly required by a provision of that implementing regulation, comparing data on the expression of a genetically modified plant, generated from plants treated with the intended herbicide, with data generated from parental plants which have not been treated with the intended herbicide is a useful measure which does not conflict, as such, with the text of Implementing Regulation No 503/2013. The data derived from such a comparison may be regarded as supplementing the data obtained in accordance with the procedures required by Implementing Regulation No 503/2013 itself, without their use for the purposes of the comparative analysis referred to in that regulation amounting to an infringement of Section 1.3.1 of Part II of Annex II thereto.

52      Moreover, as regards the applicant’s argument that the Commission’s interpretation of Implementing Regulation No 503/2013 would allow field trials of the relevant herbicide at ‘minute’ levels (see paragraph 34 above), suffice it to note that the doses of glyphosate applied in the present case were not minute. That argument of the applicant is therefore based on a purely speculative premiss and, consequently, it must be rejected on that ground.

53      In the third place, the applicant’s complaint that, in essence, the Commission failed to take due account of the data stemming from the Zanatta et al. (2020) study, must also be rejected as unfounded.

54      It is common ground that the authors of the Zanatta et al. (2020) study analysed the impact of glyphosate on the metabolism of plants and that the data obtained from that study were relevant as regards the impact of high doses of glyphosate on the gene expression of the modified soybean. However, as is apparent from Sections 3.5.6 and 3.5.7 of EFSA’s opinion of 25 September 2019, in the context of an overall analysis of the significant differences between the modified soybean and the non-genetically modified comparator, which took into account the potential impact on plant metabolism and the natural variability observed for the set of non-genetically modified reference varieties, the GMO Scientific Panel had concluded that, taking into account the natural variability observed for the set of non-genetically modified reference varieties, none of the differences identified in the agronomic and phenotypic characteristics tested between the modified soybean and the non-genetically modified comparator required further assessment of their potential environmental impact.

55      Where it claims, in essence, that the Commission failed to take due account of the data from the Zanatta et al. (2020) study, the applicant fails to link that study to the overall analysis of the data referred to in paragraphs 3.5.6 and 3.5.7 of EFSA’s opinion of 25 September 2019. It is therefore not clear to what extent the data from that specific study were such as to modify the data obtained in other studies referred to by the GMO Scientific Panel in its overall analysis. In those circumstances, it is in no way possible to understand to what extent the Zanatta et al. (2020) study could have affected the Scientific Panel’s conclusion in Section 3.5.7 of EFSA’s opinion of 25 September 2019.

56      What is more, it should be borne in mind that, as the Commission rightly pointed out, glyphosate had been applied to young leaves of plants for the purposes of conducting the Zanatta et al. (2020) study. However, young leaves of plants are in no way representative of what is used for food or feed. Relying on the Zanatta et al. (2020) study, the applicant expresses concerns which, at most, have a connection with the characteristics that the modified soybean may possess at a growth stage which takes place prior to the growth stage which is of interest in the context of the application of Regulations No 1829/2003 and No 530/2013, namely the moment at which a genetically modified organism is used for food and feed.

57      Moreover, the applicant’s argument that the authors of the Zanatta et al. (2020) study did not attach any importance to the fact that the intended herbicide had been applied to young leaves of plants and that there was no plausible explanation as to why that factor had any impact on the relevance of the study must be rejected. The mere fact that the authors of that study did not explain further the impact of the material chosen for carrying out that study could not prevent EFSA and, subsequently, the Commission from determining, on their own initiative, to what extent the conclusions of that study were relevant for the modified soybean.

58      Furthermore, and finally, in so far as the applicant suggests that the Commission was wrong to confine itself to referring to EFSA’s opinions, without, however, formulating its own assessments, it must be pointed out that the Commission is not prevented from partially or wholly adopting the assessments set out in an opinion of EFSA, but nor is it required to reproduce those assessments or to replace them on each occasion with its own reasoning. Where there is no information which seriously casts doubt on EFSA’s opinions, the scientific analysis contained in such an opinion may, in principle, be sufficient to enable the Commission to adopt a decision, without there even being any need for additional scientific assessments by the Commission (see, to that effect and by analogy, judgment of 7 March 2019, Sweden v Commission, T‑837/16, EU:T:2019:144, paragraph 67).

59      In the fourth place, when the applicant submits that the Commission disregarded certain publications cited in the request for internal review and that was unjustified, since those publications clearly established that there was a correlation between the level of the doses of glyphosate applied (including a high level), the agronomic performance and the composition of the plants, it must be stated that it is impossible for the Court to know to which publications the applicant actually refers. The accuracy of the applicant’s assertions cannot therefore be verified in any way.

60      Admittedly, it may be inferred from the evidence in the file that the applicant’s assertions relate to the fifth paragraph on page 7 of the detailed assessment, which, in turn, is also very imprecise and reads as follows:

‘Regarding the other publications you cite, the experimental data mentioned in those publications are considered important evidence of potential impact of herbicide treatment on plant metabolism. However, they are not directly relevant to this [genetically modified] stack soybean because they were generated using material produced in a glyphosate-free environment or using material not produced from a specific event or they were generated from material not produced from seeds/grain.’

61      However, the applicant does not dispute, in a substantiated manner, the reasons given by the Commission in that fifth paragraph. The application does not include any specific argument specifically calling into question the finding that the data contained in those publications were not ‘directly relevant to this [genetically modified] stack soybean because they were generated using material produced in a glyphosate-free environment or using material not produced from a specific event or they were generated from material not produced from seeds/grain’.

62      In the light of the foregoing, that is to say, in the absence of the applicant having demonstrated a manifest error of assessment or an error of law on the part of the Commission, in terms of Article 4(1)(a) and Article 16(1)(a) of Regulation No 1367/2006, read in conjunction with Article 5(a) of Implementing Regulation No 503/2013 and Sections 1.2.2.3 and 1.3.1 and Section 1.3.2.1(b) of Annex II, Part II, thereto, the first part of the single plea in law must be rejected as unfounded.

 The second part, concerning the assessment of allergenicity and toxicity

63      The applicant submits, in essence, that the responses given by the Commission in the detailed assessment in relation to the concerns expressed in the request for internal review concerning the assessment of the toxicity and allergenicity of the modified soybean are vitiated by manifest errors of assessment and amount to an infringement of Article 4(1)(a) and Article 16(1)(a) of Regulation No 1367/2006, read in conjunction with Article 5(a) of Implementing Regulation No 503/2013, and with Section 1.4(a) and (c), Section 1.4.1(a) and Sections 1.4.4.1 and 1.5 of Part II of Annex II to Implementing Regulation No 503/2013. The Commission accepted data from the field trials carried out which were wholly inadequate for assessing potential immunogenic and toxic impacts of the components of the modified soybean on humans in the course of normal consumption. In so doing, the Commission failed to give any adequate consideration to the potential for toxicity or allergenicity in the modified soybean.

64      Thus, in the first place and more specifically, the Commission infringed Section 1.4 of Part II of Annex II to Implementing Regulation No 503/2013, which concerns the study of the toxicity of the changes arising from the genetic modification of the genetically modified plant under study.

65      First, contrary to what follows from Section 1.4.1(a) of Part II of Annex II to Implementing Regulation No 503/2013, the Commission merely analysed each protein newly expressed by the genetically modified plant under study in isolation. Indeed, the Commission should have determined holistically, namely in tandem with other components of the modified soybean, whether the newly expressed proteins represented a specific risk, whether allergenic or adjuvantic, or a non-specific risk, such as chronic inflammation, to the immune system.

66      As regards the specific risk analysis, the Commission examined the safety of the various Bt toxins, in particular toxins Cry1Ac and Cry1A.105, separately from one another and in isolation from the effect of certain components present in the plant, such as protease inhibitors. That approach is incorrect. There is an obvious potential for synergistic effects between inhibitors of protease, which naturally occurs in soybean, on the one hand, and the Bt toxins, the expression of which results from the genetic modification of the modified soybean (toxins Cry1Ac and Cry1A.105), on the other. The combination of the two is likely to lead to a delay in the degradation of the Bt toxins in the gut after consumption. That delay in degradation extends the exposure of the intestinal immune system to the Bt toxins, and that prolonged exposure may trigger or enhance the toxic effects of the Bt toxins and may trigger or enhance chronic inflammation or other immune responses. In order to refute the concerns linked to that issue which had been raised in the request for internal review, the Commission referred to EFSA’s opinion; however, that body had also made errors of a scientific nature. EFSA considered, in essence, that the modified soybean did not raise any human safety concerns so far as concerns Bt toxins. In fact, the data on which EFSA relied in order to assess the acute toxicity of Bt toxins were data obtained from studies which did not take account of synergistic effects in the stack, namely feeding studies with the isolated toxins in mice. In addition, those studies examined the potential toxic effects of Bt toxins in the absence of protease inhibitors. In the present case, only experiments which capture the interaction (the synergistic effects) between Bt toxins and protease inhibitors could demonstrate whether, and to what extent, human health may be impacted by consumption of the modified soybean.

67      Finally, the Commission completely ignored the fact that interactions between the characteristics of the modified soybean could trigger non-specific immune reactions, such as chronic inflammation. The Commission should also have analysed the chronic inflammation potential in detail.

68      Second, according to the applicant, it is to be expected that the modified soybean will be exposed repeatedly to higher doses of glyphosate, which may lead to a higher quantity of glyphosate residues or glyphosate metabolites in the harvest. Herbicide residues in harvests have potential effects on the microbiome. Chronic exposure to food or feed produced from plants to which high quantities of glyphosate have been applied could lead to significant changes in the intestinal bacterial flora. Those changes can also impact the immunogenicity of Bt toxins. However, the Commission did not request that animal feeding studies be conducted to verify the effects of glyphosate residues or glyphosate metabolites on the microbiome. It follows that the Commission infringed Section 1.4(a) of Part II of Annex II to Implementing Regulation No 503/2013.

69      Third, by concluding in Section 1.3.1 of the detailed assessment that no hypothesis was identified which required a feeding study concerning the modified soybean to be carried out the Commission misapplied the second paragraph of Section 1.4.4.1 of Part II of Annex II to Implementing Regulation No 503/2013. The 90-day feeding studies which were performed with parental plants were insufficiently targeted to enable the toxicity or immunogenicity hypotheses put forward by the applicant to be verified and which ought to have been considered in the context of the testing of the modified soybean. The applicant maintains that it has established the requisite hypotheses with regard to the potential synergistic effects resulting from the combination of events in the stack together with the naturally occurring protease inhibitors and the exposure to high amounts of herbicide. The potential for synergistic effects in the modified soybean are so obvious that, in order to conclude that the stack is safe, the Commission should have requested animal feeding tests of the modified soybean which would make it possible to understand the combinatorial, synergistic or accumulated effects caused by the stacking of traits in the modified soybean.

70      Fourth, the applicant refers to Section 1.4(c) of Part II of Annex II to Implementing Regulation No 503/2013, according to which the applicant must ‘identify potential adverse effects of new constituents and determine their highest dose level that does not result in adverse effects’. However, the ‘highest dose level’ cannot be determined by testing Bt proteins in isolation from the other constituents of the modified soybean. The Commission therefore also infringed Section 1.4(c) of Part II of Annex II to Implementing Regulation No 503/2013.

71      In the second place, the Commission’s analysis of the allergenicity of the modified soybean is also flawed, which amounts to an infringement of Section 1.5 of Annex II to Part II of Implementing Regulation 503/2013.

72      First, the Commission erred in its assessment of a publication referred to in the request for internal review, namely the Santos-Vigil et al. (2018) article. That publication reports on the allergenic effects of Bt toxins when ingested. It is true that, in support of its view, the Commission relied on an external report on immunogenicity which was commissioned by EFSA, namely the Parenti et al. (2019) study. However, the conclusion drawn by the Commission from that external report is incorrect. In actual fact, far from rectifying the errors made by the Commission and EFSA, that external report rather supports the applicant’s hypothesis that Bt toxins can trigger immune reactions.

73      Admittedly, the Parenti et al. (2019) study concludes that the dosage of Bt toxins expressed in food plants is too low to trigger those allergenic effects in humans when consumed. However, the Parenti et al. study (2019) fails to take into account the enhancing effect that protease inhibitors could have on Bt toxins.

74      Moreover, the Parenti et al. (2019) study emphasises that the microbiome may influence the immunogenicity of Bt toxins after ingestion. In view of the fact that glyphosate shows antibiotic activity impacting the composition of the microbiome, it could be concluded that changes caused by glyphosate in the intestinal bacteria can also have an impact on the immunogenicity of Bt toxins. The only appropriate way of assessing the effect of the toxins, taken together with the impact of the ‘PI proteins’ and any impact of glyphosate exposure on the microbiome was, in the present case, to simulate ingestion of the entire stacked event, that is to say by means of animal feeding tests of the modified soybean.

75      Second, the Commission suggested that EFSA had assessed the allergenicity of the modified soybean as a whole and that there was no evidence that the genetic modification was likely substantially to change the overall allergenicity of the modified soybean compared to its non-genetically modified comparators and to the non-genetically modified reference varieties tested. However, that claim is categorical. The applicant notes that it had stated that neither the Commission nor EFSA were in a position properly to conclude that the newly expressed proteins were safe when combined with each other or with the other components of the modified soybean, including the protease inhibitors.

76      The Commission contests the applicant’s arguments.

77      Where the applicant disputes the Commission’s considerations on toxicity, it must be considered that it refers to Section 1.3 of the detailed assessment. By contrast, it refers necessarily to Section 1.4 of that detailed assessment when it calls into question the Commission’s observations on allergenicity.

78      In the first place, the applicant’s arguments concerning the toxicity of the modified soybean are unconvincing.

79      First, by its objections, the applicant suggests that the Commission examined the newly expressed proteins in the modified soybean in isolation from the other (natural) components of that plant, in particular separately from protease inhibitors, and that the Commission did not analyse the non-specific risks of those newly expressed proteins, thereby infringing the third sentence of Section 1.4.1(a) of Part II of Annex II to Implementing Regulation No 503/2013. According to that provision, the ‘potential interaction [of newly expressed proteins] with other plant constituents [must be] evaluated’ by the applicant for authorisation.

80      In Section 1.3.2 of the detailed assessment, the Commission provided a response to the doubts raised in the request for internal review as to the failure to examine the ‘potential interaction’ between the newly expressed proteins of the modified soybean and the natural components of that plant, such as protease inhibitors. It is apparent from Section 1.3.2 that EFSA had already assessed the publications referred to by the applicant in its request for internal review. It is also apparent from Section 1.3.2 that the results of EFSA’s assessment had been summarised in the document of the working group of the GMO Scientific Panel entitled ‘Scientific advice on the internal review under Regulation (EC) No 1367/2006 of the Commission’s decision authorising the placing on the market of genetically modified maize MON 87427 x MON 89034 x 1507 x MON 88017 x 59122 and subcombinations’. According to the authors of that document, the conclusions of the studies cited by the applicant, in its request for internal review, in support of the hypothesis concerning the existence of increased toxicity of Bt proteins, linked to the presence of protease inhibitors in soybean, had not given rise to concerns for human and animal health. Moreover, it is apparent from the minutes of the 122nd meeting of the working group of the GMO Scientific Panel that the latter had rejected the data from the MacIntosh et al. (1990) and Mesén-Porras (2020) studies for the simple reason that they could not be ‘extrapolated to infer potential adverse effects of specific [genetically modified] soybeans expressing Cry proteins on species other than insects’. According to the working group of the GMO Scientific Panel, ‘Bt proteins act on the midgut of larvae of susceptible insect species via a presumed mode of action that is not relevant for humans and animals (other than insects)’ and ‘the gastrointestinal tract of mammals, including humans, lacks receptors with high specific affinity for these proteins, and no safety concerns related to the toxicity or allergenicity of [genetically modified] plants expressing Bt proteins have been identified in those [genetically modified] plants assessed by the [GMO Scientific Panel]’. It was on the basis of those assessments that the same working group finally considered that the MacIntosh et al. (1990) and Mesén-Porras (2020) studies did not give rise to concerns for human and animal health.

81      In the application, the applicant fails to address that reasoning. It makes no reference to the observations in Section 1.3.2 of the detailed assessment as such. The arguments put forward in the application concerning that interaction are merely a repetition of the arguments put forward in the request for internal review, the purpose of which was to cast doubt on the content of the authorisation decision.

82      It follows that the arguments referred to in paragraphs 65 and 66 above are inadmissible. The arguments and pleas directed against the administrative act in respect of which an internal review was sought are inadmissible since the applicant is not entitled in his or her own right, under Article 263(2) and (4) TFEU, to challenge an authorisation decision directly before the Courts of the European Union (see, to that effect, judgment of 4 April 2019, ClientEarth v Commission, T‑108/17, EU:T:2019:215, paragraphs 26 to 28).

83      The applicant’s arguments referred to in paragraphs 65 and 66 above are, in any event, also unfounded. The applicant has not shown that the contested decision is vitiated by a manifest error of assessment. It does not even attempt to explain why the conclusion that the examination of studies such as the MacIntosh et al. (1990) and Mesén-Porras (2020) studies did not give rise to serious doubts, in the present case, since the mode of action of Bt proteins on the digestive system of insects is not relevant for humans and animals other than insects, is insufficient or inaccurate. In particular, it does not explain how the mode of action of Bt proteins can have an impact on the gastrointestinal tract of humans and animals other than insects if one starts from the principle – which, moreover, is not disputed by the applicant, in a substantiated manner and with supporting evidence – that humans and animals are not (or are not sufficiently) equipped with receptors with high specific affinity for those proteins.

84      It follows that the applicant’s arguments concerning the infringement of Section 1.4.1(a) of Part II of Annex II to Implementing Regulation No 503/2013, on account of the Commission’s failure to examine the specific risks posed by the modified soybean from the perspective of the ‘potential interaction’ between the newly expressed proteins and the protease inhibitors naturally present in soybean, must be rejected.

85      Next, according to the applicant, the Commission ignored the fact that the interactions between the characteristics of the modified soybean could trigger non-specific immune reactions, such as chronic inflammation, and did not analyse chronic inflammation potential in detail (see paragraph 67 above). There is no need to determine whether that argument relates to toxicity and therefore has a sufficiently direct relationship with Section 1.4 of Part II of Annex II to Implementing Regulation No 503/2013, whether it is related to allergenicity and, therefore, to Section 1.5 of the first paragraph of Part II of Annex II to that regulation, or whether it relates to both toxicity and allergenicity, as the applicant submitted in response to a question from the Court.

86      In any event, the applicant’s argument referred to in paragraph 85 above is unfounded. The applicant has not successfully shown that the examination carried out by the Commission was insufficient as regards the chronic inflammation potential of the modified soybean. Furthermore, it must be stated that EFSA’s analysis in the technical report, in particular in Section 2.3.3 thereof, and on which the Commission’s examination is based, is not, in fact, insufficient. It is common ground that the Parenti et al. (2019) study concludes that the dosage of Bt toxins expressed in food plants is too low to trigger allergenic effects in humans. Admittedly, the authors of that study do not appear to have taken account of the potential enhancing effects which protease inhibitors may have on Bt toxins. However, EFSA, for its part, analysed those toxins, at least in terms of the Kunitz trypsin protease inhibitor, without having been able to establish the existence of adverse effects on human or animal health. In the light of those factors, the mere invocation of the chronic inflammation potential of the modified soybean does not render EFSA’s scientific considerations implausible, with the result that no manifest error of assessment on the part of the Commission can be found.

87      Second, the applicant submits, in essence, that the Commission infringed Section 1.4(a) of Part II of Annex II to Implementing Regulation No 503/2013 by failing to request that animal feeding tests be carried out to verify the effects of glyphosate residues or metabolites of glyphosate on the microbiome, despite the fact that it is to be expected that the modified soybean will be exposed to higher doses of glyphosate and that herbicide residues in harvests have potential effects on intestinal bacteria flora (see paragraph 68 above).

88      However, as the Commission notes, the authorisation procedure for genetically modified food and feed, laid down in Articles 4 and 16 of Regulation No 1829/2003, does not cover the assessment of the potential effects of pesticide residues on human health, including possible cumulative effects. Moreover, that conclusion is reflected in paragraphs 106 and 107 of the judgment of 12 September 2019, TestBioTech and Others v Commission (C‑82/17 P, EU:C:2019:719). The assessment requested by the applicant does not fall within the scope of Regulation No 1829/2003 and Implementing Regulation No 503/2013, but within the scope of Regulation (EC) No 396/2005 of the European Parliament and of the Council of 23 February 2005 on maximum residue levels of pesticides in or on food and feed of plant and animal origin and amending Council Directive 91/414/EEC (OJ 2005 L 70, p. 1).

89      It follows that an infringement of Section 1.4(a) of Part II of Annex II to Implementing Regulation No 503/2013 cannot be found and that the applicant’s argument referred to in paragraph 68 above must be rejected as unfounded.

90      Third, the applicant complains that the Commission merely relied on 90-day feeding tests of parental plants in animals, without, however, requesting the submission of data obtained from feeding tests of the modified soybean itself, despite the fact that there are obvious potential synergistic effects in the modified soybean (see paragraph 69 above).

91      In that regard, it should be noted that, according to the first paragraph of Section 1.4.4.1 of Part II of Annex II to Implementing Regulation No 503/2013, the obligation to carry out 90-day feeding studies with genetically modified whole food and feed in rodents only exists for the assessment of food and feed containing, consisting of or produced from genetically modified plants with a single transformation event or with stacked transformation events which are not obtained by conventional crossing of genetically modified plants containing a single transformation event.

92      The modified soybean is a product that has been obtained by conventional crossing of genetically modified plants containing single transformation events (see paragraphs 3 and 4 above). Therefore, the modified soybean does not fall within the scope of the first paragraph of Section 1.4.4.1 of Part II of Annex II to Implementing Regulation No 503/2013, which means that that provision cannot have been infringed by the Commission.

93      In any event, in view of the fact that the modified soybean is a genetically modified organism for the purposes of the second sentence of the second paragraph of Section 1.4.4.1 of Part II of Annex II to Implementing Regulation No 503/2013, EFSA had, in its opinion of 25 September 2019, stated in essence that the molecular characterisation, the comparative analysis and the toxicological assessment of parental plants MON 87751, MON 87701, MON 87708 and MON 89788 (the four single events) had not led to relevant findings concerning the stability and expression of the inserts or the interaction between the transformation events. Relevant toxicological concerns had also not been identified. EFSA reached that conclusion only after an application had been made to Monsanto for the submission of additional studies concerning parental plants MON 87708, MON 87701 and MON 89788.

94      The applicant does not explain how further studies on rodents, concerning the modified soybean as a whole, were still necessary on the date of adoption of the contested decision, even though it had already been found that there had been no relevant information concerning the stability, expression of the inserts and interaction between the transformation events. The applicant’s argument referred to in paragraph 69 above must therefore be rejected as unfounded.

95      Fourth, the applicant’s complaint that, in essence, the applicant for authorisation could not have determined the ‘highest dose level that does not result in adverse effects’ of the new constituents of the modified soybean, since Bt proteins were assessed in isolation from the other constituents of that plant (see paragraph 70 above), cannot succeed either.

96      Arguments seeking to attribute errors to the applicant for authorisation in its application are inadmissible. It has been held that, in the case of an action brought against a decision on a request for review, only arguments and pleas seeking to demonstrate possible errors of law or assessment made by the Commission in the decision on the request for internal review, such as the contested decision, are admissible, as opposed to arguments concerning possible errors made by the applicant in the application for authorisation, which are, by contrast, inadmissible (see, to that effect, judgments of 6 October 2021, ClientEarth v Commission, C‑458/19 P, EU:C:2021:802, paragraph 49, and of 4 April 2019, ClientEarth v Commission, T‑108/17, EU:T:2019:215, paragraphs 53, 233 and 235). Therefore, Section 1.4(c) of Part II of Annex II to Implementing Regulation No 503/2013 cannot be found to have been infringed in the present case.

97      The applicant’s argument referred to in paragraph 70 above is, in any event, unfounded. The reference to Section 1.4(c) of Part II of Annex II to Implementing Regulation No 503/2013 is put forward in support of the applicant’s argument that EFSA and the Commission should have required further 90-day feeding studies of the modified soybean in animals to be carried out due to the potential interactions between the Bt proteins and other plant components. According to the second sentence of the second paragraph of Section 1.4.4.1 of Part II of Annex II to Implementing Regulation No 503/2013, further 90-day feeding studies with the genetically modified food and feed in rodents are, however, to be carried out only ‘where indications of potential adverse effects are identified during the assessment of: (i) the stability of the inserts; (ii) the expression of the inserts; and (iii) the potential synergistic or antagonistic effects resulting from the combination of the transformation events’. At the end of its scientific analysis based in particular on the information provided by Monsanto, EFSA had not identified any problems as regards the stability of the inserts. Moreover, it is apparent from Section 3.6.3.4 of EFSA’s opinion of 25 September 2019 that Monsanto had submitted four studies concerning 90-day feeding tests with the genetically modified food and feed in rodents for each of the single transformation events (MON 87751, MON 87701, MON 87708 and MON 89788), which is perfectly in line with the second paragraph of Section 1.4.4 of Part II of Annex II to Regulation No 503/2013. Those studies are the studies listed in Table 1 at page 8 of EFSA’s opinion of 25 September 2019. According to EFSA, no adverse effects were established for any of those four events. Furthermore, EFSA evaluated a series of additional information requested from Monsanto on the 90-day studies on genetically modified food and feed MON 87708 and MON 87701, and a new study on genetically modified food and feed stemming from MON 89788, as is also apparent from Section 3.6.3.4 of EFSA’s opinion of 3 July 2019. Following that assessment, EFSA concluded that the administration of the highest dose tested did not have adverse effects. The applicant does not link its argument concerning the failure to identify the ‘highest dose level that does not result in adverse effects’ to EFSA’s complex scientific assessment, which the Commission endorsed. Accordingly, it does not render the Commission’s assessments in that regard implausible and, therefore, a manifest error of assessment has not been demonstrated.

98      In the second place, it is necessary to reject the applicant’s arguments concerning the allergenicity of the modified soybean, by which it alleges, in essence, an infringement of Section 1.5 of Part II of Annex II to Implementing Regulation No 503/2013. The answers given by the Commission to the questions raised in the request for internal review on this issue are set out in Section 1.4 of the detailed assessment.

99      In Section 1.4 of the detailed assessment, the Commission recalled, in essence, that, in the request for internal review, the applicant had criticised the fact that the allergenicity of the Bt Cry1Ac protein, that is to say, the source of the Cry1A.105 protein, present in the modified soybean, had not been analysed in detail. In support of its reasoning, the applicant had referred to the Santos-Virgil et al. (2018) study, according to which the Cry1Ac protein is considered allergenic. The Commission responded to that criticism of the applicant by referring to the Parenti et al. study (2019). The authors of that study had been instructed by EFSA to assess the adjuvanticity and immunogenicity of certain proteins in question. According to the Commission, it is apparent from that study that the immunogenicity of Cry proteins was plausible. However, it also followed that the allergenic effect did not pose a problem, since the dosage of Bt toxins, expressed in food plants, was too low.

100    The applicant disputes those assessments, stating that the Parenti et al. (2019) study did not take into account the enhancing effect that protease inhibitors could have on Bt toxins. However, in so doing, it does not demonstrate the existence of a manifest error of assessment vitiating the contested decision, since it does not place the Commission’s line of argument in the correct context, namely the context surrounding the assessment referred to in the Santos-Virgil et al. (2018) study.

101    Admittedly, the Parenti et al. (2019) study, which served as the basis for the Commission’s response and which appears in Annex A.19 to the application, does not appear to refer to protease inhibitors, even though, in the second sentence of the fifth paragraph of point 3.1.2 of that publication, it is stated as follows:

‘Other aspects of protein structure likely to be relevant for adjuvanticity and immunogenicity are solubility, stability, size, and the compactness of the overall fold. These aspects reflect dependency of allergenicity on transport over mucosal barriers and susceptibility to proteases. Even if most allergens can be grouped into a small number of structural classes, a specific fold characterising allergenic protein was not identified, and few, if any, structural features are currently known to be common for allergens, which rather shows [a] wide variety of secondary structure and folding.’

102    However, the results of the Parenti et al. study (2019) and protease inhibitors were assessed by EFSA regardless. It is apparent from line 71 of the technical report that:

‘In relation to potential immunogenic effects of Bt proteins, EFSA previously published comprehensive scientific reports addressing similar questions on the EFSA assessment of [good manufacturing practice] and the potential effects of Bt proteins on the immune system (e.g. [the GMO Scientific Panel], 2017b; EFSA, 2018b; Parenti et al., 2019). Briefly, the [GMO Scientific Panel] does not find indications that Bt proteins in the stacked [genetically modified] maize and stacked [genetically modified] soybean assessed by [that Scientific] Panel might act as adjuvants with the potential to enhance a specific immunoglobulin E (IgE) response and to favour the development of an allergic reaction. Furthermore, because none of the newly expressed proteins in the assessed [genetically modified] plants showed potential for allergenicity considering current knowledge, no reasons for concerns are expected regarding the simultaneous presence of these newly expressed proteins in the stacked [genetically modified] maize and stacked [genetically modified] soybean.’

103    The authors of the technical report also took into account certain protease inhibitors. Thus, it is stated in that report that:

‘Finally, the assessment of allergenicity of the whole [genetically modified] plants was also considered. Protease inhibitors are compounds naturally occurring in specific crops. In such respect, it is noted that composition of the [genetically modified] plant was also analysed, and it included an analysis of the Kunitz trypsin inhibitor in the case of soybean. Considering all the information available, the [GMO Scientific Panel] considers that there is no evidence that the genetic modification might substantially change the overall allergenicity of the stacked [genetically modified] maize and stacked [genetically modified] soybean assessed when compared to the [non-genetically modified] comparators and the [non-genetically modified] reference varieties tested.’

104    It should be borne in mind that the technical report is at the heart of the detailed assessment and of the contested decision, as is apparent from that decision. Consequently, contrary to what the applicant suggests, when it rejected the relevance of the results of the Santos-Virgil et al. (2018) study, the Commission did not in any way confine itself to merely taking into account the Parenti et al. (2019) study. In fact, it had also adopted the considerations set out in that report, according to which, in essence, the allergenicity of the modified soybean had well and truly been analysed in terms of protease inhibitors, in particular, and, at least, in terms of the Kunitz trypsin protease inhibitor.

105    The applicant does not address the results of the Santos-Virgil et al. (2018) and Parenti et al. (2019) studies in the context of EFSA’s conclusions in the technical report, referred to in paragraph 102 above. In particular, it fails to address the issue of allergenicity of the modified soybean taking into account the data on the Kunitz trypsin protease inhibitor. It follows that the applicant does not consider the issue of the effect of those protease inhibitors on the Bt toxins in the light of all the information on which the Commission’s detailed assessment was based. It therefore does not succeed in rendering implausible the assessments contained in the detailed assessment in that regard, which is why it does not succeed in demonstrating the existence of a manifest error of assessment on the part of the Commission. Its arguments must therefore be rejected as unfounded.

106    Finally, in so far as, in the context of allergenicity, the applicant complains that the Commission did not request Monsanto to submit feeding studies of the modified soybean, that argument cannot succeed either.

107    First, it is only in the context of toxicity that such studies may be requested from the applicant for authorisation, as is apparent in particular from the provisions of Section 1.4.4.1 of Part II of Annex I, subparagraph (d) of the second paragraph of Section 1.4, subparagraph (e) of the third paragraph of Section 1.4.1, and Sections 1.4.3, 1.4.4 and 1.4.4.1 of Part II of Annex II to Implementing Regulation No 503/2013.

108    Second, and in any event, the applicant does not explain why further feeding studies of the modified soybean as a whole to rodents were still necessary, when EFSA had already stated in the technical report that ‘there [was] no evidence that the genetic modification might substantially change the overall allergenicity of the stacked [genetically modified] maize and stacked [genetically modified] soybean’.

109    Furthermore, the applicant doubts whether the Commission assessed the allergenicity of the modified soybean as a whole and whether there was sufficient evidence to conclude that the newly expressed proteins were safe when combined with each other or with the other components of the modified soybean, including protease inhibitors; the related claims in the detailed assessment are categorical (see paragraph 75 above).

110    Those doubts, which were not expressed as a specific criticism, but which read as an overall conclusion, are unfounded. In Annex A to the technical report, EFSA carried out a thorough scientific analysis of the arguments raised in the request for internal review as regards the allergenicity of the modified soybean, including with regard to the impact of protease inhibitors. In the light of that analysis, which the Commission fully endorsed, the Commission’s conclusion cannot be described as ‘categorical’.

111    In view of the foregoing, the second part of the single plea must be rejected in its entirety, as must that plea as such.

112    It follows that the present action must be rejected as unfounded.

 Costs

113    Under Article 134(1) of the Rules of Procedure of the General Court, the unsuccessful party is to be ordered to pay the costs if they have been applied for in the successful party’s pleadings. Since the applicant has been unsuccessful, it must be ordered to bear its own costs and to pay those incurred by the Commission, in accordance with the form of order sought by the latter.

On those grounds,

THE GENERAL COURT (Second Chamber)

hereby:

1.      Dismisses the action;

2.      Orders TestBioTech eV to bear its own costs and to pay those incurred by the European Commission.

Marcoulli Tomljenović Valasidis

Delivered in open court in Luxembourg on 18 October 2023.

V. Di Bucci M. van der Woude

Registrar President

Table of contents

Background to the dispute

Forms of order sought

Law

Preliminary observations

The single plea in law

The first part, concerning exposure to the application of herbicides

The second part, concerning the assessment of allergenicity and toxicity

Costs

*      Language of the case: English.