On 2015 Sep 01, Zhiping Pang commented:

We appreciate that Dr. Rinaman acknowledges that our conclusions are consistent with previous studies (Alhadeff et al., 2012; Dickson et al., 2012; Dossat et al., 2011; Skibicka, 2013), but strongly disagree with her surmise that our study is flawed based on the specificity of the mouse line used to manipulate GLP-1 neurons. We apologize that, primarily due to space limitations, we did not cite all the papers from Dr. Rinaman and colleagues. However, we argue that the specificity of the mouse line may be not as clearly doubted as Dr. Rinaman states, nor do we believe that our conclusions depend on that specificity alone. Based on the totality of our experimental work, we believe that our conclusions, as stated in the paper, are sound. As with all published scientific work, we provide experimental evidence for a particular hypothesis that is logical and plausible, but do not claim that our model provides a definitive answer to the question — in this study, how central GLP-1 regulates feeding. Thus, we feel that the comment from Dr. Rinaman et al. is much more apodictic and definitive than the phrasing of our paper’s conclusions.

We would like to respond to the specific concerns raised by Dr. Rinaman and her co-authors in the comment posted on PubMed Commons.

Dr. Rinaman et al. suggested that we claimed that Phox2b is GLP-1 specific:

We did not state, explicitly or implicitly, that endogenous expression of Phox2b and GLP-1 are 100% overlapping. We acknowledged that it is possible that not all GLP-1 expressing neurons express the phox-2b-cre transgene and that other types of neurons may express Phox-2b-cre as well. The purpose of utilizing the Phox2b-Cre transgenic line was to assess whether a defined group of central GLP-1 neurons was involved in regulating food intake. Our experimental results provide evidence that GLP-1 neurons likely participate in the regulation of food intake, although they do not exclude the involvement of non-GLP-1 Phox2b-Cre expressing neurons. Specifically, our data support a specific role of GLP-1 neurons in the regulation of food intake behavior in the following ways: a) the anorexic effects induced by the activation of Phox2b-Cre expressing neurons are blocked by the GLP-1R specific blocker Exendin-9 (also discussed below); b) retrograde-labeled NTS-VTA projecting neurons are positive for GLP-1; c) Cre-activated expression of EYFP colocalizes with GLP-1 in brain sections detected by a commercially available antibody (Peninsula Laboratories T-4363) (Zheng and Rinaman, 2015; Zheng et al., 2015); d) injection of CNO at the VTA in DREADD-expressing animals leads to suppressed food intake after 5 hours. In ongoing, unpublished studies, we have expressed Cre-dependent channelrhodopsin in NTS neurons of Phox2b-Cre transgenic mice to express channelrhodopsin in Phox2b-Cre positive NTS neurons, and found that neuronal activation by optical stimulation of the NTS nerve terminals is blocked by Exendin-9. This presents additional evidence to support that GLP-1R (GLP-1 receptor) is expressed in Phox2b-cre expressing cells. Taken together, these findings, along with reports from Scott et al (Scott et al., 2011) and the collection of studies cited throughout our manuscript, lead us to propose evidence for the involvement of GLP-1 signaling in the VTA in the regulation of feeding behavior.

Additionally, the transgenic mice used in this study were created based on the Bacteria Artificial Chromosome (BAC) technology. Unlike in the case of gene knockins generated by homologous recombination, in the case of transgenics, including BAC transgenics (Heintz, 2001), the introduced foreign gene is randomly inserted into the genome (Beil et al., 2012) and the expression of the transgene is influenced by epigenetic factors and genetic background (Chan et al., 2012). Therefore, the expression of the transgene does not always faithfully mimic endogenous gene expression. Indeed, the three Phox2b-Cre transgenic lines generated in Dr. Elmquist’s laboratory exhibit different expression patterns (Scott et al., 2011). Given these considerations, we did not conclude that Phox2b-Cre was only expressed in GLP-1 neurons or vice-versa. As described in the paper, the Phox2b-Crehese animals were employed as a tool to interrogate the function of a group of GLP-1 expressing neurons in regulating food intake behavior.

Due to size limitation, full response please refer to (please copy the hyperlink address): https://www.dropbox.com/s/0n129f2ugn3tgjz/Response.pdf?dl=0

On 2015 Aug 24, Linda Rinaman commented:

Phox2b is Not Specifically Expressed by Hindbrain GLP-1 Neurons

M.R. Hayes, University of Pennsylvania, Philadelphia, PA, USA; L. Rinaman, University of Pittsburgh, Pittsburgh, PA, USA; K.P. Skibicka, Sahlgrenska Academy at the University of Gothenburg, Sweden; S.Trapp, University College London, London, U.K.; D.L. Williams, Florida State University, Tallahassee, FL, USA

The first part of this study sought to extend published work [1-7] supporting a role for central GLP-1 signaling in suppressing palatable food intake. For this purpose, the authors virally expressed DREADDs within the caudal medulla of transgenic Cre-driver line mice, followed by chemogenetic DREADD activation to increase or decrease the activity of transfected neurons. Unfortunately, their experimental design depends on a Phox2b-Cre mouse model [8] that is non-specific for GLP-1 neurons.

Phox2b is expressed by a diverse set of autonomic-related neurons distributed throughout the nucleus of the solitary tract (NTS), area postrema (AP), dorsal motor nucleus of the vagus (DMV), and other regions [9-13], including catecholaminergic and HSD-2-positive neurons that innervate mesolimbic, hypothalamic, and other central targets [14-16]. The present study includes a supplementary figure (S1) purporting to show co-localization of GLP-1 immunolabeling with mCherry reporter expression, but the depicted coronal section is well rostral to the established location of GLP-1 neurons in rats and mice [17-19].  Thus, the GLP-1 immunolabeling is non-specific, and the authors present no credible evidence that GLP-1 neurons express virally-encoded DREADDs.

It is not surprising that food intake was suppressed in mice in which Phox2b-expressing AP, DMV, and NTS neurons were transfected to express hM3Dq. CNO activation of these neurons should disrupt physiological functions and activate stress-sensitive GLP-1 neurons [19-21] whether or not they express DREADDs. However, other than food intake, no physiological or behavioral measures were performed.  The authors report that the hypophagic effect of CNO was specific to the high-fat diet, with no effect on chow intake, but their experimental design and results are insufficient to support this claim.  Further, i.p. injection of Exendin-9 was reported to block the hypophagic effect of CNO (Figure 1F). The basis for this effect is difficult to understand, because the utilized i.p. dose of Exendin-9 is well below the established threshold for antagonizing GLP-1 receptors in the periphery, let alone within the brain [22,23].  In addition, stereotaxic injections were used to deliver a GLP-1 receptor agonist or CNO into the ventral midbrain of mice just before measuring their food intake (Figure 2), with no consideration of how acute surgery, presumably in anesthetized mice, might affect subsequent feeding behavior.   

In summary, we believe that the present report is seriously flawed.  Although the authors' conclusions are consistent with previous work in rats, their report fails to demonstrate a specific role for endogenous central GLP-1 signaling in the control of palatable food intake in mice.

REFERENCES

1. Dossat, A.M., et al., J Neurosci, 2011. 31: 14453.

2. Alhadeff, A.L., L.E. Rupprecht, and M.R. Hayes, Endocrinol, 2012. 153: 647.

3. Mietlicki-Baase, E.G., et al., Am J Physiol Endocrinol Metab, 2013. 305: E1367.

4. Dickson, S.L., et al., J Neurosci, 2012. 32: 4812.

5. Skibicka, K.P., Front Neurosci, 2013. 7: 181.

6. Richard, J.E., et al., Plos One, 2015. 10(3).

7. Hayes, M.R., L. Bradley, and H.J. Grill, Endocrinol, 2009. 150: 2654.

8. Scott, M.M., et al., J Clin Invest, 2011. 12: 2413.

9. Kang, B.J., et al., J Comp Neurol, 2007. 503: 627.

10. Lazarenko, R.M., et al., J Comp Neurol, 2009. 517: 69.

11. Geerling, J.C., P.C. Chimenti, and A.D. Loewy, Brain Res, 2008. 1226: 82.

12. Mastitskaya, S., et al., Cardiovasc Res, 2012. 95: 487.

13. Brunet, J.F. and A. Pattyn, Curr Opin Genet Dev, 2002. 12: 435.

14. Geerling, J.C. and A.D. Loewy, J Comp Neurol, 2006. 497: 223.

15. Delfs, J.M., et al., Brain Res, 1998. 806: 127.

16. Mejias-Aponte, C.A., C. Drouin, and G. Aston-Jones, J Neurosci, 2009. 29: 3613.

17. Llewellyn-Smith, I.J., et al., Neurosci, 2011. 180: 111.

18. Vrang, N., et al., Brain Res, 2007. 1149: 118.

19. Rinaman, L., Am J Physiol, 1999. 277: R582.

20. Maniscalco, J.W., A.D. Kreisler, and L. Rinaman, Front Neurosci, 2013. 6: 199.

21. Maniscalco, J.W., et al., J Neurosci, 2015. 35: 10701.

22. Williams, D.L., D.G. Baskin, and M.W. Schwartz, Endocrinol, 2009. 150: 1680.

23. Kanoski, S.E., et al., Endocrinol, 2011. 152: 3103.

On 2015 Aug 24, Linda Rinaman commented:

Phox2b is Not Specifically Expressed by Hindbrain GLP-1 Neurons

M.R. Hayes, University of Pennsylvania, Philadelphia, PA, USA; L. Rinaman, University of Pittsburgh, Pittsburgh, PA, USA; K.P. Skibicka, Sahlgrenska Academy at the University of Gothenburg, Sweden; S.Trapp, University College London, London, U.K.; D.L. Williams, Florida State University, Tallahassee, FL, USA

The first part of this study sought to extend published work [1-7] supporting a role for central GLP-1 signaling in suppressing palatable food intake. For this purpose, the authors virally expressed DREADDs within the caudal medulla of transgenic Cre-driver line mice, followed by chemogenetic DREADD activation to increase or decrease the activity of transfected neurons. Unfortunately, their experimental design depends on a Phox2b-Cre mouse model [8] that is non-specific for GLP-1 neurons.

Phox2b is expressed by a diverse set of autonomic-related neurons distributed throughout the nucleus of the solitary tract (NTS), area postrema (AP), dorsal motor nucleus of the vagus (DMV), and other regions [9-13], including catecholaminergic and HSD-2-positive neurons that innervate mesolimbic, hypothalamic, and other central targets [14-16]. The present study includes a supplementary figure (S1) purporting to show co-localization of GLP-1 immunolabeling with mCherry reporter expression, but the depicted coronal section is well rostral to the established location of GLP-1 neurons in rats and mice [17-19].  Thus, the GLP-1 immunolabeling is non-specific, and the authors present no credible evidence that GLP-1 neurons express virally-encoded DREADDs.

It is not surprising that food intake was suppressed in mice in which Phox2b-expressing AP, DMV, and NTS neurons were transfected to express hM3Dq. CNO activation of these neurons should disrupt physiological functions and activate stress-sensitive GLP-1 neurons [19-21] whether or not they express DREADDs. However, other than food intake, no physiological or behavioral measures were performed.  The authors report that the hypophagic effect of CNO was specific to the high-fat diet, with no effect on chow intake, but their experimental design and results are insufficient to support this claim.  Further, i.p. injection of Exendin-9 was reported to block the hypophagic effect of CNO (Figure 1F). The basis for this effect is difficult to understand, because the utilized i.p. dose of Exendin-9 is well below the established threshold for antagonizing GLP-1 receptors in the periphery, let alone within the brain [22,23].  In addition, stereotaxic injections were used to deliver a GLP-1 receptor agonist or CNO into the ventral midbrain of mice just before measuring their food intake (Figure 2), with no consideration of how acute surgery, presumably in anesthetized mice, might affect subsequent feeding behavior.   

In summary, we believe that the present report is seriously flawed.  Although the authors' conclusions are consistent with previous work in rats, their report fails to demonstrate a specific role for endogenous central GLP-1 signaling in the control of palatable food intake in mice.

REFERENCES

1. Dossat, A.M., et al., J Neurosci, 2011. 31: 14453.

2. Alhadeff, A.L., L.E. Rupprecht, and M.R. Hayes, Endocrinol, 2012. 153: 647.

3. Mietlicki-Baase, E.G., et al., Am J Physiol Endocrinol Metab, 2013. 305: E1367.

4. Dickson, S.L., et al., J Neurosci, 2012. 32: 4812.

5. Skibicka, K.P., Front Neurosci, 2013. 7: 181.

6. Richard, J.E., et al., Plos One, 2015. 10(3).

7. Hayes, M.R., L. Bradley, and H.J. Grill, Endocrinol, 2009. 150: 2654.

8. Scott, M.M., et al., J Clin Invest, 2011. 12: 2413.

9. Kang, B.J., et al., J Comp Neurol, 2007. 503: 627.

10. Lazarenko, R.M., et al., J Comp Neurol, 2009. 517: 69.

11. Geerling, J.C., P.C. Chimenti, and A.D. Loewy, Brain Res, 2008. 1226: 82.

12. Mastitskaya, S., et al., Cardiovasc Res, 2012. 95: 487.

13. Brunet, J.F. and A. Pattyn, Curr Opin Genet Dev, 2002. 12: 435.

14. Geerling, J.C. and A.D. Loewy, J Comp Neurol, 2006. 497: 223.

15. Delfs, J.M., et al., Brain Res, 1998. 806: 127.

16. Mejias-Aponte, C.A., C. Drouin, and G. Aston-Jones, J Neurosci, 2009. 29: 3613.

17. Llewellyn-Smith, I.J., et al., Neurosci, 2011. 180: 111.

18. Vrang, N., et al., Brain Res, 2007. 1149: 118.

19. Rinaman, L., Am J Physiol, 1999. 277: R582.

20. Maniscalco, J.W., A.D. Kreisler, and L. Rinaman, Front Neurosci, 2013. 6: 199.

21. Maniscalco, J.W., et al., J Neurosci, 2015. 35: 10701.

22. Williams, D.L., D.G. Baskin, and M.W. Schwartz, Endocrinol, 2009. 150: 1680.

23. Kanoski, S.E., et al., Endocrinol, 2011. 152: 3103.

On 2015 Sep 01, Zhiping Pang commented:

We appreciate that Dr. Rinaman acknowledges that our conclusions are consistent with previous studies (Alhadeff et al., 2012; Dickson et al., 2012; Dossat et al., 2011; Skibicka, 2013), but strongly disagree with her surmise that our study is flawed based on the specificity of the mouse line used to manipulate GLP-1 neurons. We apologize that, primarily due to space limitations, we did not cite all the papers from Dr. Rinaman and colleagues. However, we argue that the specificity of the mouse line may be not as clearly doubted as Dr. Rinaman states, nor do we believe that our conclusions depend on that specificity alone. Based on the totality of our experimental work, we believe that our conclusions, as stated in the paper, are sound. As with all published scientific work, we provide experimental evidence for a particular hypothesis that is logical and plausible, but do not claim that our model provides a definitive answer to the question — in this study, how central GLP-1 regulates feeding. Thus, we feel that the comment from Dr. Rinaman et al. is much more apodictic and definitive than the phrasing of our paper’s conclusions.

We would like to respond to the specific concerns raised by Dr. Rinaman and her co-authors in the comment posted on PubMed Commons.

Dr. Rinaman et al. suggested that we claimed that Phox2b is GLP-1 specific:

We did not state, explicitly or implicitly, that endogenous expression of Phox2b and GLP-1 are 100% overlapping. We acknowledged that it is possible that not all GLP-1 expressing neurons express the phox-2b-cre transgene and that other types of neurons may express Phox-2b-cre as well. The purpose of utilizing the Phox2b-Cre transgenic line was to assess whether a defined group of central GLP-1 neurons was involved in regulating food intake. Our experimental results provide evidence that GLP-1 neurons likely participate in the regulation of food intake, although they do not exclude the involvement of non-GLP-1 Phox2b-Cre expressing neurons. Specifically, our data support a specific role of GLP-1 neurons in the regulation of food intake behavior in the following ways: a) the anorexic effects induced by the activation of Phox2b-Cre expressing neurons are blocked by the GLP-1R specific blocker Exendin-9 (also discussed below); b) retrograde-labeled NTS-VTA projecting neurons are positive for GLP-1; c) Cre-activated expression of EYFP colocalizes with GLP-1 in brain sections detected by a commercially available antibody (Peninsula Laboratories T-4363) (Zheng and Rinaman, 2015; Zheng et al., 2015); d) injection of CNO at the VTA in DREADD-expressing animals leads to suppressed food intake after 5 hours. In ongoing, unpublished studies, we have expressed Cre-dependent channelrhodopsin in NTS neurons of Phox2b-Cre transgenic mice to express channelrhodopsin in Phox2b-Cre positive NTS neurons, and found that neuronal activation by optical stimulation of the NTS nerve terminals is blocked by Exendin-9. This presents additional evidence to support that GLP-1R (GLP-1 receptor) is expressed in Phox2b-cre expressing cells. Taken together, these findings, along with reports from Scott et al (Scott et al., 2011) and the collection of studies cited throughout our manuscript, lead us to propose evidence for the involvement of GLP-1 signaling in the VTA in the regulation of feeding behavior.

Additionally, the transgenic mice used in this study were created based on the Bacteria Artificial Chromosome (BAC) technology. Unlike in the case of gene knockins generated by homologous recombination, in the case of transgenics, including BAC transgenics (Heintz, 2001), the introduced foreign gene is randomly inserted into the genome (Beil et al., 2012) and the expression of the transgene is influenced by epigenetic factors and genetic background (Chan et al., 2012). Therefore, the expression of the transgene does not always faithfully mimic endogenous gene expression. Indeed, the three Phox2b-Cre transgenic lines generated in Dr. Elmquist’s laboratory exhibit different expression patterns (Scott et al., 2011). Given these considerations, we did not conclude that Phox2b-Cre was only expressed in GLP-1 neurons or vice-versa. As described in the paper, the Phox2b-Crehese animals were employed as a tool to interrogate the function of a group of GLP-1 expressing neurons in regulating food intake behavior.

Due to size limitation, full response please refer to (please copy the hyperlink address): https://www.dropbox.com/s/0n129f2ugn3tgjz/Response.pdf?dl=0