Drug Responsiveness in Patient-Derived Rectal Organoids Correlates with Clinical Response in CF Subjects: A Real-Life Analysis

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper describes the use of 2D intestinal organoids in CF, with measurements with the Ussing chamber and measures responses to CFTR modulators and wants to compare these results with clinical responses in the corresponding patients with CF.

The use of 2D organoids (and measurements with the Ussing chamber) is interesting as they can be used in organoids with no and residual CFTR function. 

However, this study/manuscript needs to be adapted. The results, measurements and statistics need to become scientifically sound and furthermore they need to be written down in a structured way, without grammatical and spelling mistakes.

Some important comments:

r 19 Why is chosen for the abbreviation OG? This seems not very logical

r 21 trikafta > change this to ETI

r 87-89 rewrite this phrase, is not a logical sentence right now

r 88 rewrite luma/Ivacaftor to Lumacaftor/Ivacaftor

r 88 change to 'at least homozygous F508del'

figure 1 what is meant by vehicle, please explain

figure 1 why is an unpaired student's t-test used and not a paired one?

figure 1 S11: how many CF patients are used in this group?

figure 2 can you explain the increase of SCC, while there is no response in the organoids? is this due to variability in the SCC?

Are measurements in organoids dome in triplicate? Please mention.

r 117 what is mentioned by +/- 29?

r 138-140 rephrase please, and check grammar

r 146 figure 3A is mentioned as there is only figure 3?

r 146-150 what is the message here? The sentence is too long, and given information is not clear. Furthermore, what is meant by 'data of FEV1 would not be technically suitable for the analysis'? Why do you represent them in figure 2, are these data technically suitable?

r 161 Should be 'matched with', but maybe 'correlated with' is a better naming. But actually I do not agree with this conclusion, as the results of the 2D organoids may correlate with the other CFTR function biomarker Sweat Chloride Concentration (SCC), but they do not correlate with the FEV1, at least this could not be analyzed (see above mentioned remark). So this conclusion needs to be adapted.

 

 

Author Response

We thank the reviewers for the useful suggestions that helped us to significantly improve our study, please find below our response to a point-by-point basis. We also added an Addendum with technical details.

Reviewer 1

This paper describes the use of 2D intestinal organoids in CF, with measurements with the Ussing chamber and measures responses to CFTR modulators and wants to compare these results with clinical responses in the corresponding patients with CF.

The use of 2D organoids (and measurements with the Ussing chamber) is interesting as they can be used in organoids with no and residual CFTR function. 

However, this study/manuscript needs to be adapted. The results, measurements and statistics need to become scientifically sound and furthermore they need to be written down in a structured way, without grammatical and spelling mistakes.

Some important comments:

Comment 1: r 19 Why is chosen for the abbreviation OG? This seems not very logical

Response 1: We agree with the reviewer. We apologize for our typing mistake. The abbreviation 'OG' was corrected to 'PDROs'.

Comment 2: r 21 trikafta > change this to ETI

Response 2: Done.

Comment 3: r 87-89 rewrite this phrase, is not a logical sentence right now

Response 3: Done. We rewrote this part, hoping that this version is easier to comprehend. Thank you.

Comment 4: r 88 rewrite luma/Ivacaftor to Lumacaftor/Ivacaftor

Response 4: Done.

Comment 5: r 88 change to 'at least homozygous F508del'

Response 5: Done.

Comment 6: figure 1 what is meant by vehicle, please explain

Response 6: Vehicle means the use of the substance DMSO, dimethyl sulfoxide, which is used to solubilize small-molecule drugs. We change it on the text accordingly. Thanks

Comment 7: figure 1 why is an unpaired student's t-test used and not a paired one?

 

Response 7: Since the samples have different sizes, we used a t-test for independent samples instead of a t-test for paired samples, maintaining this choice - for consistency - even in the few cases where the samples to be compared have the same size.

However, we verified that, based on the data available to us, the use of a t-test for independent samples reproduced the same result as a t-test for paired samples and that there were no significant reductions in the power of the test.

Comment 8: figure 1 S11: how many CF patients are used in this group?

Response 8: In total 10 CF patients were evaluated, in which four were also characterized in vitro during this study.  S11 is our positive control (patient harboring F508del)

Comment 9: figure 2 can you explain the increase of SCC, while there is no response in the organoids? is this due to variability in the SCC?

Response 9: There is some variability in SSC and increased values of SCC are not consistent with a response to CFTR modulators. The limited increase of SCC shown in the middle of figure 2 is consistent with the negligible response in organoids of the same subject shown in the graph on the right in the same figure

Comment 10: Are measurements in organoids dome in triplicate? Please mention.

Response 10: In the graphic, each point represents a single electrophysiological measurement. Hence, organoids were measured minimally in triplicate. This information was added to the legend of the graphic.

Comment 11: r 117 what is mentioned by +/- 29?

Response 11: It refers to the standard deviation of the average -30.1 mmol/L found for SSC in the analyzed group.

Comment 12: r 138-140 rephrase please, and check grammar.

Response 12: Done. Thanks.

Comment 13: r 146 figure 3A is mentioned as there is only figure 3?

Response 13: Thanks. We corrected it.

Comment 14: r 146-150 what is the message here? The sentence is too long, and given information is not clear. Furthermore, what is meant by 'data of FEV1 would not be technically suitable for the analysis'? Why do you represent them in figure 2, are these data technically suitable?

Response 14: Thank you for helping us to make our points clearer to the reader. 

The issue here is that data do not satisfy three of the principal assumptions which justify the use of linear regression models, such as linearity and additivity of the relationship between dependent and independent variables, homoscedasticity (constant variance) of the errors, and normality of error distribution).

Indeed, if any of these assumptions are violated, then the forecasts, confidence intervals, and scientific insights yielded by a regression model may be (at best) inefficient or (at worst) seriously biased or misleading.  Consequently, for FEV1 data, we could not perform the same correlation analysis we used for Delta SC.

Conversely, the same data can be analyzed as a group, as shown in figure 2, where we wondered whether the treatment modifies FEV1 when a control group was matched with the corresponding treated group. In this case a statistically significant difference is measured when an appropriate type of analysis such as Wilcoxon matched-pairs signed rank test is utilized.

We rewrote this part in the main text. Hoping it is clear now.

PLEASE SEE ADDENDUM  for details.

Comment 15: r 161 Should be 'matched with', but maybe 'correlated with' is a better naming. But actually I do not agree with this conclusion, as the results of the 2D organoids may correlate with the other CFTR function biomarker Sweat Chloride Concentration (SCC), but they do not correlate with the FEV1, at least this could not be analyzed (see above mentioned remark). So this conclusion needs to be adapted.

Response 15: Thank you. We modified the sentence specifying that we refer to ST as for the technical reasons discussed we could not perform a similar analysis for FEV1.

ADDENDUM

 

If the size of the samples whose means are being compared is different, it is not possible to use the paired t-test.

 

Below we show the statistical analyses carried out as follows.

 

- if the samples have different sizes (but greater than 5), we use the t-test for paired data, removing the unpaired data, and then the t-test for unpaired samples. We calculate the reference threshold of the p-value by applying Bonferroni's correction to the 5% threshold level, thus obtaining 0.025 as the new reference value.

 

- If the samples are of equal size (but greater than 5), we use the t-test for paired data.

 

- if one or both samples (before or after removing paired data) we also use a Wilcoxon test, which, being non-parametric, can also be applied in these cases, bearing in mind, however, that as the sample size decreases, so does the power of the test.

 

 

S1
W57G / A234D
Vehicle	ETI
3,79	62,52
8,52	24,64
5,67	30,3
2,85	32,18
5,67	29,33
4,73	17,97
15,15

 

Paired t-test (unpaired data eliminated)

data:  dataset_51[, 1] and dataset_51[, 2]

t = -4.1411, df = 5, p-value = 0.004494

alternative hypothesis: true mean difference is less than 0

95 percent confidence interval: -Inf -14.17924

sample estimates: mean difference -27.61833

 

Welch Two Sample t-test

data:  dataset_51_x and dataset_51_y

t = -4.038, df = 5.6261, p-value = 0.003896

alternative hypothesis: true difference in means is less than 0

95 percent confidence interval: -Inf -13.43909

sample estimates: mean of x ; mean of y: 6.625714; 32.823333

 

Results: The tests remain significant even after applying Bonferroni's correction, which brings the p-value threshold to 0.025.

 

S2
R74W+V201M+D1270N / CFTRdele22-2434D
Vehicle	ETI
3,03	7
1,52	5,85
0,94	10,61
0,76	28,79
2,09
3,76
1,58

 

 

Paired t-test (unpaired data removed)

data:  dataset_S2[, 1] and dataset_S2[, 2]

t = -2.0311, df = 3, p-value = 0.0676

alternative hypothesis: true mean difference is less than 0

95 percent confidence interval: -Inf, 1.824755

sample estimates: mean difference -11.5

 

Welch Two Sample t-test

data:  dataset_S2_x and dataset_S2_y

t = -2.0741, df = 3.0362, p-value = 0.06432

alternative hypothesis: true difference in means is less than 0

95 percent confidence interval: -Inf ,1.434685

sample estimates: mean of x mean of y : 1.954286, 13.062500

 

Results: both tests indicate no statistically significant difference between the mean of the two samples.

 

 

 

S3
F508del / 711+5A
Vehicle	ETI
0,55	14,18
1,12	13,24
1,7	19,48
	47,7

 

Paired t-test (unpaired data removed)

data:  dataset_S3[, 1] and dataset_S3[, 2]

t = -8.5751, df = 2, p-value = 0.006664

alternative hypothesis: true mean difference is less than 0

95 percent confidence interval: -Inf, -9.569066

sample estimates: mean difference -14.51

 

Welch Two Sample t-test

data:  dataset_S3_x and dataset_S3_y

t = -2.7673, df = 3.01, p-value = 0.03473

alternative hypothesis: true difference in means is less than 0

95 percent confidence interval: -Inf, -3.395611

sample estimates: mean of x mean of y: 1.123333,  23.650000

 

Wilcoxon rank sum exact test

data:  dataset_S3_x and dataset_S3_y

W = 0, p-value = 0.02857

alternative hypothesis: true location shift is less than 0

 

Results: the most reliable result is that provided by the Wilcoxon test, given that the sample size is less than 5 (a value below which t-tests lose significant power).

 

 

 

S4
A559T / A559T
Vehicle	ETI
0,97	16,09
0,58	17,03
0,94	7,21
1,12	12,67

 

Paired t-test

data:  dataset[, 1] and dataset[, 2]

t = -5.4283, df = 3, p-value = 0.006135

alternative hypothesis: true mean difference is less than 0

95 percent confidence interval: -Inf -6.994384

sample estimates: mean difference: -12.3475

 

Wilcoxon rank sum exact test

data:  dataset_x and dataset_y

W = 0, p-value = 0.01429

alternative hypothesis: true location shift is less than 0

 

Results: the most reliable result is that provided by the Wilcoxon test, given that the sample size is less than 5 (a value below which t-tests lose significant power).

 

 

S5
R347P / R347P
Vehicle	ETI
2,27	6,24
2,45	8,52
2,09	5,85
3,21	15,12
2,64	7
2,27	6,24

 

Paired t-test

data:  dataset[, 1] and dataset[, 2]

t = -4.384, df = 5, p-value = 0.003564

alternative hypothesis: true mean difference is less than 0

95 percent confidence interval: -Inf, -3.065636

sample estimates: mean difference: -5.673333

 

Results: we applied the paired sample t-test as the two samples have the same size.

 

 

 

 

 

S6
L227R / L227R
Vehicle	ETI
0,94	0,39
1,7	3,24
1,12	1,55
0,58	0,18
0,21	1,33
0,76	1,52
1,15	2,24
0,76	1,52
0,21	1,12
	1,52

 

 

Paired t-test (unpaired values removed)

data:  dataset[, 1] and dataset[, 2]

t = -2.7073, df = 8, p-value = 0.01339

alternative hypothesis: true mean difference is less than 0

95 percent confidence interval: -Inf, -0.1969197

sample estimates: mean difference -0.6288889

 

Welch Two Sample t-test

data:  dataset_x and dataset_y

t = -5.5531, df = 3.0159, p-value = 0.005678

alternative hypothesis: true difference in means is less than 0

95 percent confidence interval: -Inf -7.125976

sample estimates: mean of x mean of y: 0.9025,   13.2500

 

Results: the Bonferroni correction brings the p-values threshold to 0.025. The paired t-test shows a statistically significant difference between the mean of the two samples, while the unpaired t-test does not. However, given that the samples have two different sizes, the result of the unpaired t-test is more trustable and informative.

 

S7
F508del / dupl. exon1-3
Vehicle	ETI
2,45	43,18
4,55	41,67
11,73	18,91
60,42	64,21
55,48	70,64
	18,91

 

 

Paired t-test (unpaired values removed)

data:  dataset[, 1] and dataset[, 2]

t = -2.7187, df = 4, p-value = 0.02653

alternative hypothesis: true mean difference is less than 0

95 percent confidence interval: -Inf, -4.489105

sample estimates: mean difference: -20.796

 

 

Welch Two Sample t-test

data:  dataset_x and dataset_y

t = -1.0269, df = 7.4227, p-value = 0.1684

alternative hypothesis: true difference in means is less than 0

95 percent confidence interval: -Inf 13.26444

sample estimates: mean of x mean of y: 26.926    42.920

 

Results: the Bonferroni correction brings the p-values threshold to 0.025. The paired t-test still shows a statistically significant difference between the mean of the two samples, while the unpaired t-test does not. However, given that the samples have two different sizes, the result of the unpaired t-test is more trustable and informative.

 

 

 

S8
R1162X / 3849+10kBC>T
Vehicle	ETI
1,7	4,92
3,8	10,8
0,1	6,8

 

Paired t-test

data:  dataset[, 1] and dataset[, 2]

t = -4.6493, df = 2, p-value = 0.02164

alternative hypothesis: true mean difference is less than 0

95 percent confidence interval: -Inf, -2.097779

sample estimates: mean difference  -5.64

 

Wilcoxon rank sum exact test

data:  dataset_x and dataset_y

W = 0, p-value = 0.05

alternative hypothesis: true location shift is less than 0

 

Results: the Bonferroni correction brings the p-values threshold to 0.025. The paired t-test shows a statistically significant difference between the mean of the two samples, while Wilcoxon test does not.

 

 

S9
2183AA>G / N1303K
Vehicle	ETI
0,76	0,91
0,58	1,33
0,39	1,3
0,97

 

Paired t-test (unpaired data removed)

data:  dataset[, 1] and dataset[, 2]

t = -2.6082, df = 2, p-value = 0.06046

alternative hypothesis: true mean difference is less than 0

95 percent confidence interval: -Inf 0.07213142

sample estimates: mean difference: -0.6033333

 

Welch Two Sample t-test

data:  dataset_x and dataset_y

t = -2.7519, df = 4.6053, p-value = 0.02197

alternative hypothesis: true difference in means is less than 0

95 percent confidence interval: -Inf, -0.1281055

sample estimates: mean of x mean of y: 0.675     1.180

 

Wilcoxon rank sum exact test

data:  dataset_x and dataset_y

W = 1, p-value = 0.05714

alternative hypothesis: true location shift is less than 0

 

Results: the Bonferroni correction brings the p-values threshold to 0.025 (paired t-test is not considered, since it is not applicable for samples of different sizes).

 

 

 

 

S10
R553X / 2789+5G>A
Vehicle	ETI
1,88	3,61
1,88	3,21
3,97	4,73
6,42	9,48
2,09	22,91
2,27
0,18
0
0,18

 

 

Paired t-test (unpaired data removed)

data:  dataset[, 1] and dataset[, 2]

t = -1.4432, df = 4, p-value = 0.1112

alternative hypothesis: true mean difference is less than 0

95 percent confidence interval: -Inf, 2.64359

sample estimates: mean difference -5.54

 

 

Welch Two Sample t-test

data:  dataset_x and dataset_y

t = -1.7764, df = 4.2782, p-value = 0.07282

alternative hypothesis: true difference in means is less than 0

95 percent confidence interval: -Inf 1.19106

sample estimates: mean of x mean of y: 2.096667  8.788000

 

Results: given that the samples have two different sizes, the result of the unpaired t-test is more trustable and informative.

 

 

S11
F508del / F508del
Vehicle	ETI
4,55	42,97
1,12	34,27
4,55	39,76
1,12	42,97
3,58	33,33

 

 

Paired t-test

data:  dataset[, 1] and dataset[, 2]

t = -17.07, df = 4, p-value = 3.454e-05

alternative hypothesis: true mean difference is less than 0

95 percent confidence interval: -Inf, -31.22041

sample estimates: mean difference -35.676

 

 

 

Reviewer 2 Report

Comments and Suggestions for Authors

This is a solid study with minor concerns. The work provides valuable real-world validation of organoid-based theratyping for CFTR modulators, though the novelty is somewhat limited.

I have only several minor concerns:

Abstract: 

Line 22: I think is more appropriate not to use the commercial name of the drug (Trikafta), but only its components (ETI). 

Introduction:

Line 36: EMA has changed the indications of ETI, as the authors correctly report in the discussion (lines 250 - 252). It would be more appropriate the description of chronology of indications in EU. 

Discussion: 

There are published data on L227R and A559T responsiveness to ETI (in vitro Bihler et al, JCF 23 (2024) 664–675, and in vivo Burgel PR, Lancet Respir Med. 2024 Nov;12(11):888-900). I think that these evidence should be cited. 

Line 222 - 225: I agree completely with this statement. To strenghten this reasoning, I suggest the reference to the data by French researchers (Burgel PR et al.,  Lancet Respir Med. 2025 Sep 3:S2213-2600(25)00190-0. ).

Author Response

We thank the reviewers for the useful suggestions that helped us to significantly improve our study, please find below our response to a point-by-point basis. We also added an Addendum with technical details.

 

Reviewer 2

This is a solid study with minor concerns. The work provides valuable real-world validation of organoid-based theratyping for CFTR modulators, though the novelty is somewhat limited.

I have only several minor concerns:

Abstract: 

Comment 1: Line 22: I think is more appropriate not to use the commercial name of the drug (Trikafta), but only its components (ETI). 

Response 1: We modified it accordingly. Thanks.

Introduction:

Comment 2: Line 36: EMA has changed the indications of ETI, as the authors correctly report in the discussion (lines 250 - 252). It would be more appropriate the description of chronology of indications in EU. 

 Response 2: Thanks. We modified this part as suggested.

Comment 3: Discussion: 

There are published data on L227R and A559T responsiveness to ETI (in vitro Bihler et al, JCF 23 (2024) 664–675, and in vivo Burgel PR, Lancet Respir Med. 2024 Nov;12(11):888-900). I think that these evidence should be cited. 

Response 3: We added the suggested references in the most appropriate location. Thanks

Comment 4: Line 222 - 225: I agree completely with this statement. To strenghten this reasoning, I suggest the reference to the data by French researchers (Burgel PR et al.,  Lancet Respir Med. 2025 Sep 3:S2213-2600(25)00190-0. ).

Response 4: We added the suggested references. Thanks