Gd³⁺ Complexes Conjugated to Cyclodextrins: Hydroxyl Functions Influence the Relaxation Properties

Round 1

Reviewer 1 Report

see attached file

Comments for author File: Comments.pdf

Author Response

Responses to reviewers

Reviewer 1

1) For a better understanding of the experiments, the experimental section, which now is included in supporting material, should be moved to the main text

Response: As the experimental section has 17 pages, it seemed easier to separate the results from the experiments. However if the Editor wants only one document we can reorganize the text.

2) Abstract, line 22: “various hydroxylated and permethylated 21 b-CDs via an amide linker” should be “….via an amide bond”

Response: modification done in the text

3) Figure 1: Gd-TTHA derivatives are erroneously labelled as compound 4 should have R1, R2= Me and compound 5 should have R1, R2=H

Response: Yes, it is a mistake, the correction has been done in the graph and in the text.

4) Line 210: “HTBU” is likely “HBTU”

Response: modification done in the text

5) Remove the concentrations from legends of figures 2 and 3 as the data are reported as relaxivities (r1/mM-1s-1) thus normalized to 1 mM concentrations. The concentrations of the solutions which were used for NMRD profiles acquisition should be indicated in the experimental part.

Response: Done

6) After complexation of the ligands with Gd(III), the absence of free Gd ions should be checked…especially with cyclodextrins containing systems which could retain metal ions weakly bound to OH groups

Response: The absence of free Gd(III) ions was systematically checked by a xylenol orange test on each sample. The following sentence has been added in the experimental part: “The absence of free Ln³⁺ was systematically checked on each sample using the xylenol orange test. The Gd³⁺ concentrations were also checked systematically by ICP-MS and/or BMS measurements.”

7) Supporting info, figure 2: in the legend of the figure it is written that the spectrum has been acquired in CD3CN while in the text it is stated that it was acquired in CDCl3

Response: We thanks the referee for this remark, CD₃CN was used, the text has been modified in the supporting info.

8) Relevant information (i.e. on the reorientational correlation time) could be obtained through the fitting of NMRD profiles, which is missing. For example, the authors assume that the size of the different systems is the same and thus the relaxivity is not influenced by this parameter. However, the presence or the absence of 14 Me groups (difference in MW of ca. 200 D) could have an influence on the global reorientational correlation time and fitting of NMRD profiles would highlight this kind of differences.

Response: Indeed, the referee is right, a difference of molecular weight has an influence on the rotational correlation time and on the relaxivity. However the presence of the Me groups induces a ca. 300 Da difference for a total molecular weight of ca. 2000 Da. This is roughly 15% difference, and this would certainly imply a difference in the rotational correlation times which is in the typical range of errors generally obtained when rotational correlation times are determined from the fit of NMRD data. Moreover, the difference in the exchange rate will also influence the system (see below and the response to reviewer 2) therefore the fit will not yield reliable rotational correlation times for these systems, explaining why we chose not to fit the data.

9) The variable temperature 17OR2 data and high resolution NMR on Eu-complexes should be better analyzed. For macrocyclic DOTA derivatives, the presence of two isomers (corresponding to SAP and TSAP structures) characterized by different exchange dynamics has long been appreciated since the end of the nineties, and it was ascertained that the water exchange kinetics of the TSAP isomer are 50−100 ?mes faster than those of the SAP isomer. For the parent Gd-DOTA the ratio between SAP and TSAP structures is highly shifted toward the former (ca. 90:10). Here, inspection into 17O data, suggests that for Gd-DOTA-1 and Gd-DOTA-2 a higher amount of TSAP over SAP isomer, with respect to the parent Gd-DOTA, takes place. In my opinion, more information could be obtained from 1H NMR spectra of the corresponding Eu complexes simply trough a more accurate acquisition of the spectra. First of all, looking at figures 8 and 9 of supplementary, it seems that a lot of free ligand is present …the complexation should be better performed. Second, increasing the concentration of the Eu-complexes (I didn’t find the concentration of the solutions used for 1HNMR) or the number of acquisitions better spectra could be obtained which would allow for a rough estimation of the SAP/TSAP ratio. Moreover, to be thorough, 1H NMR of Eu-DOTA-3 should be included in the work. Once obtained this useful information from 1H-NMR, a complete analysis of 17O data could be done.

Response: We do not agree with the referee that there is free ligand in the solution. The complex was prepared in a 1/1 ratio and dialyzed. The peaks present in the diamagnetic region of the spectra correspond to the many protons of the cyclodextrins away from the Eu(III) centers. These peaks are dominating the spectra explaining the difficulty to obtain better data inherent to the system. The spectra were recorded at 5 mM for 24:Eu and 1 mM for 25/Eu. They were acquired with ns =2048. In short, we do not think that with these systems containing many protons in the diamagnetic range we can estimate the SAP/TSAP ratio. But we definitely agree with the referee that it could help explaining the ¹⁷O data. In order to be more specific in the manuscript, we added the following sentence:

“The SAP (square antiprismatic) and TSAP (twisted square antiprismatic) isomers of macrocyclic systems such as DOTA derivatives are known to have different water exchange rates and their ratio can be very different depending on the systems.” With the corresponding references: Aime, A. Barge, M. Botta, A. S. De Sousa and D. Parker, Angew.Chem., Int. Ed., 1998, 37, 2673; C. Platas-Iglesias, Eur. J. Inorg. Chem. 2012, 2023 – 2033

And we replaced: “Unfortunately, the ¹H NMR spectra of the Eu³⁺ analogues of 1 and 2, recorded at different temperatures, show broad resonances and did not allow distinguishing different isomers.”

By “Unfortunately, the ¹H NMR spectra of the Eu³⁺ analogues of 1 and 2, recorded at different temperatures, show broad resonances, combined with the presence of many protons of the cyclodextrins in the diamagnetic window which dominate the spectra. In overall, this prevents distinguishing different isomers.”

Author Response File: Author Response.pdf

Reviewer 2 Report

The manuscript by Bonnet, Gouhier and colleagues reports on the importance of the hydrophilicity of a ligand for the performance of its Gd-complex as MRI contrast agent. As examples the authors report the synthesis and characterization of several imaging probes based on beta-cyclodextrin as scaffold. Although the subject is not particularly innovative, the preparation is well described and the investigation techniques are appropriate to the aim of the study.

However, there are some points that need to be addressed before this work can be considered suitable for publication.

1. At Page 2, Line 53, the sentence “1 mM concentration of the probe” is not correct and should be replaced with “1 mM concentration of metal”, since a probe may contain more than one Gd-complex (e.g. multimeric agents).
2. Several times “Gd3+” is written without “3+” superscript.
3. P. 3, L. 127, there are seven 6-positions on a beta-CD, so it is more correct to say “…ligands were conjugated at one of the C-6 positions of various beta-CDs…”.
4. I have concerns about the fact that tosylation may have been achieved at only one C-6-position with simple TsCl. There are more selective procedures reported in the literature, for example by Cravotto et al. (J. Inclusion Phenom. Macrocyclic Chem. 2007, 57, 3–7) or Botta et al. (Chem. Eur. J. 2014, 20, 10944-10952). Did the authors notice (e.g. by mass spectrometry) multiple Ts- or N₃-derivatives, with consequent need of purification?
5. P. 4, L. 158, it would be nice to show a comparison between the two ¹³C NMR spectra and, even better, between the IR spectra (showing disappearance of the clearly-identifiable signal of N₃ at 2100 cm^-1).
6. “6-O-monoamino” and similar nomenclatures throughout the manuscript don’t sound correct. They should be changed, for example as “mono(6-deoxy-6-amino)” and/or as reported in the SI.
7. P. 5, L. 172, compound 17 is already peracetylated; probably the authors mean compound 7.
8. P. 4, L. 173, why now is the azide reduced by catalytic hydrogenation instead of Staudinger reaction as done previously? Later it will be again carried out with PPh₃. If there are reasons to apply either procedure to specific substrate, that should be stated.
9. P. 4, L. 177, “following by” is to be changed into “followed by”.
10. If compound 16 was obtained in good yield (70%), why did the authors replace it with compound 21? For the TTHAMA ligand the peptide coupling strategy was used, so why not for the derivative of 16?
11. P. 6, L. 192, again it would be nice to show a comparison between the three ¹³C NMR spectra of 17, 19 and 23 in the area 40-50 ppm, if the authors have them clean and not noisy.
12. P. 8, L. 235 and below, how do these results compare with Gd-TTHAMA not bound to beta-CD?
13. P. 9, L. 250 and below, the ¹H NMRD profiles (even just at 25 °C) have not been fitted, so the corresponding information mentioned in paragraph 2.3 (i.e. q, tauM and tauR) have not been calculated. This should definitely be done.
14. P. 10, L. 292 and below, the authors should add a table reporting the data (e.g. tauM) obtained by fitting the ¹⁷O NMR measurements for compounds 1-3, possibly comparing them to those in the literature for DOTA and DOTA-monoamide Gd-complexes. There should also be, at some point, a discussion about the effect on tauM of replacing a COOH of DOTA with an amide (CONHR).
15. In the SI: after Equation(1), “…1/T_2OS can are neglected…” is to be changed with “…1/T_2OS can be neglected…”.
16. In the SI: for compound 28, the assignment of the ¹³C signal at 171 ppm is wrongly assigned to CH₂COOH instead of CH₂COOH.
17. In the SI: for compound 9 the IR characterization reports a peak at 2927 cm^-1 (“CH alkyl”) and 2100 cm^-1 (“N₃”); while it is surely important that for the next compound 12 there is no more peak for the azide (see comment 6 above), the fact that the peak at 2920 cm^-1 is now assigned to “NH₂” seems a bit exaggerated. Maybe “CH alkyl + NH₂” would be more honest.
18. In the SI: could the authors observe the I and II amide bands in the IR spectrum of compound 15? If so, they have to be reported.
19. In the SI: the ¹³C NMR spectrum of compound 29 in Fig. 3 is far too noisy and not really meaningful. It should be removed and, if possible (i.e. if the sample is still available), replaced with a better one (more concentrated and/or acquired for a longer time).
20. In the SI, are the authors sure that the data at the end of the NMRD profiles (500 MHz?) are in agreement with the trend of the plot? Or maybe the profile can be cut at 80 MHz, as said in the “General Procedures” paragraph? How do these points adapt to the fitting curves (see comment 13 above)?

Author Response

Reviewer 2

1. At Page 2, Line 53, the sentence “1 mM concentration of the probe” is not correct and should be replaced with “1 mM concentration of metal”, since a probe may contain more than one Gd-complex (e.g. multimeric agents).

Response: The referee is right, the sentence has been corrected accordingly.

1. Several times “Gd3+” is written without “3+” superscript.

Response: Done

1. P. 3, L. 127, there are seven 6-positions on a beta-CD, so it is more correct to say “…ligands were conjugated at one of the C-6 positions of various beta-CDs…”.

Response: The correction has been done in the text.

1. I have concerns about the fact that tosylation may have been achieved at only one C-6-position with simple TsCl. There are more selective procedures reported in the literature, for example by Cravotto et al. (J. Inclusion Phenom. Macrocyclic Chem. 2007, 57, 3–7) or Botta et al. (Chem. Eur. J. 2014, 20, 10944-10952). Did the authors notice (e.g. by mass spectrometry) multiple Ts- or N₃-derivatives, with consequent need of purification?

Response: We change the reference 34 by Botta Chem Eur J more appropriated. The monotosylation occurs after 2 hours of stirring at room temperature. The pure product is obtained after precipitation in acetone and the structure was confirmed by HRMS and NMR.

1. 4, L. 158, it would be nice to show a comparison between the two ¹³C NMR spectra and, even better, between the IR spectra (showing disappearance of the clearly-identifiable signal of N₃ at 2100 cm^-1).

Response: A sentence has been added in the text. Unfortunately, we do not have the IR spectra anymore in our data base. HRMS and NMR confirmed the structure.

The azide reduction using Staudinger reaction led to mono(6-amino-6-deoxy)-b-cyclodextrins precursors 10-12 with yields between 40-52% which was confirmed by the disappearance of the signal in IR spectroscopy of azide function at 2199 cm^-1 and appearance of amine function at 2920 cm^-1.  

1. “6-O-monoamino” and similar nomenclatures throughout the manuscript don’t sound correct. They should be changed, for example as “mono(6-deoxy-6-amino)” and/or as reported in the SI.

Response: 6-O-monoamino has been replaced by mono(6-amino-6-deoxy) in the text.

1. P. 5, L. 172, compound 17 is already peracetylated; probably the authors mean compound 7.

Response: Yes, it is a mistake, the correction has been done in the text.

1. P. 4, L. 173, why now is the azide reduced by catalytic hydrogenation instead of Staudinger reaction as done previously? Later it will be again carried out with PPh₃. If there are reasons to apply either procedure to specific substrate, that should be stated.

Response: The Staudinger reaction has been initially performed. We tested in a second time hydrogenation to avoid the separation of phosphine oxide by product. The reactions were carried out at different times that do not correspond to the article chronology. Similar yields were obtained: 50% for compound 17 (H₂), 40-52% for compounds 10-12, and 79% for compound 18 (PPh₃).

1. P. 4, L. 177, “following by” is to be changed into “followed by”.

Response: Done

1. If compound 16 was obtained in good yield (70%), why did the authors replace it with compound 21? For the TTHAMA ligand the peptide coupling strategy was used, so why not for the derivative of 16?

Response: With the second strategy there is another step that reduce the yield from 70% to 57% (2 steps). However, the reagent DO3A 22 is a commercially available which is not the case for the synthesized reagent 13 (Scheme 1). For this reason, larger amount of compound 21 was obtained and used to immobilize the TTHA ligand.

1. P. 6, L. 192, again it would be nice to show a comparison between the three ¹³C NMR spectra of 17, 19 and 23 in the area 40-50 ppm, if the authors have them clean and not noisy.

Response: Unfortunately, we do not have a high definition spectrum for these compounds and we cannot perform another experiment.

1. P. 8, L. 235 and below, how do these results compare with Gd-TTHAMA not bound to beta-CD?

Response: The NMRD profile of Gd-TTHAMA has been reported in the reference 33.

It has been added in the text. And the following sentence has been added as well after the value 2.3 mM-1.s-1 obtained from simulation: “This is also the value reported for GdTTHA in the same conditions”

1. P. 9, L. 250 and below, the ¹H NMRD profiles (even just at 25 °C) have not been fitted, so the corresponding information mentioned in paragraph 2.3 (i.e. q, tauM and tauR) have not been calculated. This should definitely be done.

Response: We have not fitted the NMRD profiles for the DOTA derivatives as we are convinced that these systems are too complex, as it is explained in paragraphs 2.4 and 2.5, and consequently, the fit would not give physically meaningful data for the parameters. Indeed, the relaxivity values are influenced by both the second sphere relaxation effect as well as the first sphere water exchange rate, and in the fitting these two effects are difficult to separate and one can be “mathematically” compensated by the other without reflecting the reality. It is therefore difficult to propose a reliable fit and we chose not to fit those data on purpose.

We added a sentence stating that the complexity of the systems prevents a reliable fit at the end of paragraph 2.5 on page 11.

Concerning the TTHA derivatives, the mechanism will be outer sphere relaxation combined with a second sphere contribution in the case of hydroxylated cyclodextrins. Therefore as q = 0, no water exchange rate (or tM) can be obtained. In this case, the minimal distance of approach between the water molecule and Gd³⁺, and the diffusion coefficient of the molecule will be important parameters. As stated at the end of paragraph 2.3 “a purely outer sphere mechanism give a relaxivity of 2.3 mM^-1.s^-1 at 20 MHz and 25°C, in the same order of magnitude as that measured for 4”.

1. P. 10, L. 292 and below, the authors should add a table reporting the data (e.g. tauM) obtained by fitting the ¹⁷O NMR measurements for compounds 1-3, possibly comparing them to those in the literature for DOTA and DOTA-monoamide Gd-complexes. There should also be, at some point, a discussion about the effect on tauM of replacing a COOH of DOTA with an amide (CONHR).

Response: A Table (Table 1) comparing the water exchange rate of compound 3 and other DOTA derivatives has been inserted, as well as the following sentence to discuss the effect of the replacement of carboxylate functions by amide functions:

“The water exchange rate is nearly three times lower than the water exchange rate of GdDOTA (k_ex²⁹⁸ = 4.1×10⁶), and higher than that of GdDOTAM (see Table 1), which is consistent with previous observations on analogous systems. Indeed, in the case of dissociative exchange for DOTA-derivatives (which is expected here), it is generally observed that the replacement of one negatively charged carboxylate in the complex with a neutral amide decreases the water exchange rate to about one third.”

1. In the SI: after Equation(1), “…1/T_2OS can are neglected…” is to be changed with “…1/T_2OS can be neglected…”.

Response: Done

1. In the SI: for compound 28, the assignment of the ¹³C signal at 171 ppm is wrongly assigned to CH₂COOH instead of CH₂COOH.

Response: The modification has been done in the text.

1. In the SI: for compound 9 the IR characterization reports a peak at 2927 cm^-1 (“CH alkyl”) and 2100 cm^-1 (“N₃”); while it is surely important that for the next compound 12 there is no more peak for the azide (see comment 6 above), the fact that the peak at 2920 cm^-1 is now assigned to “NH₂” seems a bit exaggerated. Maybe “CH alkyl + NH₂” would be more honest.

Response: The modification has been reported in the text.

1. In the SI: could the authors observe the I and II amide bands in the IR spectrum of compound 15? If so, they have to be reported.

Response: Indeed, it is an oversight, the data have been added in the text.

In the SI: the ¹³C NMR spectrum of compound 29 in Fig. 3 is far too noisy and not really meaningful. It should be removed and, if possible (i.e. if the sample is still available), replaced with a better one (more concentrated and/or acquired for a longer time).

Response: We agree with the referee that the spectrum is not satisfactory. Unfortunately, we do not have the compound anymore to perform another experiment. The Figure 3 has been removed, and the following figures have been re-numbered consequently.

1. In the SI, are the authors sure that the data at the end of the NMRD profiles (500 MHz?) are in agreement with the trend of the plot? Or maybe the profile can be cut at 80 MHz, as said in the “General Procedures” paragraph? How do these points adapt to the fitting curves (see comment 13 above)?

Response: The same sample was used to record all the relaxivity data for a given system, thus the point measured at 400 MHz can be well included in the NMRD profile. It is normal that these values are lower than those at 80 MHz, as predicted by the SBM theory. The points at 400 MHz have been measured on a classical Bruker NMR spectrometer and not on the relaxometer. This information was missing from the experimental part, thus the following sentence has been added in the SI : “The high-field point (400 MHz) was recorded on a Bruker Nanobay 400 (9.4 T, 54.5 MHz) spectrometer in a capillary tube in H₂O and D₂O was used as an external reference to lock the spectrometer.”

Author Response File: Author Response.pdf

Reviewer 3 Report

The authors synthesized β-cyclodextrins that were amide-linked with gadolinium chelators (DOTA and TTHA derivatives), and analyzed the extent of relaxation enhancement of water 1H spins by the gadolinium ion by 1H and 17O NMR. They synthesized five types of cyclodextrins, which differed depending on whether each of the 2, 3 and 6 positions was hydroxyl or methyl groups. They found that for the TTHA ligand that directly coordinates no water molecule, the number of hydroxyl groups on the cyclodextrin correlated with relaxation enhancement, concluding that water molecules in the second sphere that hydrogen-bond to these hydroxyl groups promoted the relaxation. On the other hand, the DOTA ligands that directly coordinate water showed the results that were not as simple as above. The number of hydroxyl groups seems to have affected the water exchange rate, which complicated the mechanism of the relaxation enhancement.

Although this study cannot be said to have shown great progress as a new contrast reagent for MRI, it does show steady basic data. I have the impression that the data described will be useful for future development. I, therefore, recommend its publication.

I am not an expert in chemical synthesis, so I do not know much about the detail. However, I thought the following few points might be mistakes.

p.4, line 133

“of the 6-o-methylated beta-CD”

Isn't this inconsistent with compound 2 in Figure 1 (left)?

p.6, line 202

“confirmed the structures 21-23”

I think this may be a mistake of 24-26. The compound 22 seems to be DO3A-t-Bu ester.

p.7, line 210

“HTBU” should be “HBTU” (Hexafluorophosphate Benzotriazole Tetramethyl Uronium).

p.8, Scheme 6

Do the compounds 4 and 5 depicted here correspond to each of the compounds in Figure 1 (right)?

When swapping 4 and 5, please be careful not to contradict the contents described in the text.

Author Response

Reviewer 3

1. 4, line 133“of the 6-o-methylated beta-CD” Isn't this inconsistent with compound 2 in Figure 1 (left)?

Response: The sentence has been modified in the text for more clarity.

Original: “The GdDOTAMA complex was introduced on the small rim of the native (hydroxylated) b-CD (1), of the 6-O-methylated b-CD (2), and of the 2,3,6-O-permethylated b-CD (3). “

Modified: “The GdDOTAMA complex was introduced at one of the O-6-positions of the small rim of the native (hydroxylated) b-CD (1), of the 6-O-permethylated b-CD (2), and of the 2,3,6-O-permethylated b-CD (3). “

2. 6, line 202“confirmed the structures 21-23” I think this may be a mistake of 24-26. The compound 22 seems to be DO3A-t-Bu ester.

Response: Right, it is a mistake, it has been corrected in text.

3. 7, line 210 “HTBU” should be “HBTU” (Hexafluorophosphate Benzotriazole Tetramethyl Uronium).

Response: corrected

4. 8, Scheme 6 Do the compounds 4 and 5 depicted here correspond to each of the compounds in Figure 1 (right)? When swapping 4 and 5, please be careful not to contradict the contents described in the text.

Response: The reviewer is right, it is a mistake, it has been modified in the graph and in the text.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors properly responded to most of my comments and the manuscript can be accepted.

Reviewer 2 Report

Although the Authors have positively accepted most of the remarks, a few of them have been purposely rejected, that I consider the most important: 1) providing evidences of the NMR and IR data that the authors themselves mention in the text to prove their conclusions, and 2) the fitting of the NMRD profiles.

The visual presentation of the spectra would have strongly confirmed the statements by the Authors, that instead can still remain subject to doubts. It is also weird that the Authors claim to have lost the IR datasets, thus being unable to add the figures, while they were also able to add some missing IR data in the Experimental Part.

About the NMRD profiles, I disagree with what declared by the Authors, i.e. that fitting of such "complex systems" is difficult. In the literature there is plenty of NMRDs of more complex molecular architectures that have been properly fitted, leading to interesting and unarguable results. In fact, it doesn't make sense to acquire a whole profile without fitting it: then better just measure R1 at 20 MHz and 25 °C and calculate the relaxivity.

In my opinion, these two issues make the manuscript a bit lame; but if it is ok with the Authors to expose their work to critics by the future readers, then it is ok with me to approve the publication without further revisions.