**4. Discussion and Conclusions**

#### *4.1. HOS Inferrred from 1D and 2D Spectra*

In this work, standard NMR experiments using 1D 1H, 2D 1H-13C HSQC and 2D 1H-15N sofast HMQC pulse sequences were performed on formulated protein DPs. The NMR peak patterns from both 1D and 2D spectra are qualitatively informative for protein HOS properties, providing insight into the oligomerization of liraglutide, the HOS exchange of exenatide and the well folded HOS of calcitonin-salmon, teriparatide, insulin glargine and rituximab. In general, the 1D 1H NMR experiment provides information on the HOS profile and whether a protein is folded in formulation. Information on more specific HOS variation can be obtained from heteronuclear 2D spectra. Each 2D spectrum was sensitive to different aspect of HOS. For example, in the insulin glargine spectra, while the methyl 1H-13C spectrum showed the sidechains adopting two slowly exchanging conformers, the 1H-15N spectrum was more consistent with a single well-folded backbone conformer. The two observations were not necessarily inconsistent with each other, rather, they illustrate the complex nature of protein HOS in the formulation and the atomic level probes used by the different NMR experiments.

#### *4.2. HOS Similarity Metrics Calculated from 1D and 2D Spectra*

What is different from the pioneering work on demonstrating heteronuclear 2D NMR at protein natural abundance [41,48,49] is that the current study uses NMR on formulated DPs and also includes deriving practically achievable similarity metrics. Earlier work demonstrated the practically achievable Mahalanobis distance (DM) value of 3.3 based on the PCA of 1D 1H spectra collected on the marketed insulin reference product and follow-on products [27]. Here, we obtained the DM values of 1.95 and 3.15 using PCA and 1D 1H spectra of rituximab DPs marketed in the US and India, suggesting that a DM metrics value of less than 3.3 could be a general acceptance criterion.

While PCA can be conveniently performed on 1D spectra and has been demonstrated on 2D spectra [26,38,50], PCA is challenging to implement for 2D spectra because of the technical complications in binning the 2D spectra and avoiding non-DS peaks at the same time. An alternative method is to focus on the DS peak profile. The normalized distance comparison approach was proposed to compare 2D spectra along the axes of chemical shifts and peak intensity; however, no acceptance criteria were ever proposed [51]. Here the previous chemical shift comparison method [26] was verified using 2D 1H-13C spectra collected on insulin glargine DPs and the chemical shift different metrics (Δδ) of 4 ppb for 1H and 15 ppb for 13C were derived. Furthermore, the peak profile method [44] was adopted to compare the relative peak heights between two insulin glargine brands, where *p* values were derived from *t*-test. In these insulin spectra, 98% of the methyl cross peaks had equivalent relative peak heights between the two brands. These 2D spectral similarity metrics could be equivalent to the DM value of 1.6 obtained by using 1D spectra [27]. The methyl peak profile results represent another practically achievable similarity metrics for 2D spectral comparison.

In summary, the NMR data collected in the current study provided examples of simple experiments and analyses on formulated protein DP and demonstrated practical measurements to assess equivalence of HOS between different DPs. The metrics proposed were validated using marketed similar DPs that were manufactured differently and are proposed as a benchmark to determine the degree of similarity for protein HOS in formulated DPs.

**Supplementary Materials:** The following are available online. Tables S1–S3 and Matlab code for DM calculation.

**Author Contributions:** Conceptualization, D.A.K. and K.C.; methodology, D.W., D.A.K. and K.C.; validation, D.W. and Y.Z.; investigation, D.W., Y.Z., M.K., S.M.P. and K.C.; writing—original draft preparation, D.W., C.J.S., D.A.K. and K.C.; writing—review and editing, D.W., M.K., S.M.P., C.J.S., D.A.K. and K.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available in supplementary material.

**Acknowledgments:** This project was supported, in part, by an appointment (Y.Z. and S.M.P.) to the Research Participation Program at the CDER administered by the Oak Ridge Institute for Science and Education (ORISE) through an interagency agreement between the U.S. Department of Energy and the U.S. FDA.

**Conflicts of Interest:** The authors declare no conflict of interest.

**Disclaimer:** This article reflects the views of the author and should not be construed to represent U.S. FDA's views or policies.

**Sample Availability:** Samples of the compounds are not available from the authors.

#### **References**


MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Molecules* Editorial Office E-mail: molecules@mdpi.com www.mdpi.com/journal/molecules

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18

www.mdpi.com

ISBN 978-3-0365-2581-5