DETAILED ACTION
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.
Claims 16-19, 21, 24, 26, 27, 30-33 and 39, 41-48 and 50 are pending.
Claims 19, 21, 24, 26, 27, 30, 31, 43 and 44 have been withdrawn.
Claims 16-18, 32, 33, 39, 41, 42 and 45-48 and 50 are currently under examination.
Priority
Applicant’s claim for the benefit of a prior-filed application under 35 U.S.C. 119(e) or under 35 U.S.C. 120, 121, or 365(c) is acknowledged. Applicant has not complied with one or more conditions for receiving the benefit of an earlier filing date under 35 U.S.C. §119 as follows:
The later-filed application must be an application for a patent for an invention which is also disclosed in the prior application (the parent or original nonprovisional application or provisional application). The disclosure of the invention in the parent application and in the later-filed application must be sufficient to comply with the requirements of the first paragraph of 35 U.S.C. 112. See Transco Products, Inc. v. Performance Contracting, Inc., 38 F.3d 551, 32 USPQ2d 1077 (Fed. Cir. 1994).
The disclosure of the prior-filed application, Application No. 62/324,228, 62,345,556 and 62/458,954 fails to provide adequate support or enablement in the manner provided by the first paragraph of 35 U.S.C. 112 for one or more claims of this application.
Particularly, the provisional patent applications do not have support for dendritic or B-cells of patient-derived cells. Thus, the present claims are hereby assigned the priority date of October 18, 2018, the filing date of the present application.
There is not support in the specification as filed for the limitation “a second population of patient-derived cells comprising HLA-peptide complexes, and isolating the HLA-peptide complexes from dendritic or B-cells of the patient-derived cells
Paragraphs 11-16 of the specification recite
(a) providing a population of cells expressing a single HLA allele;
(a) providing a population of cells expressing a single HLA class I allele;
(a) providing a population of cells expressing a pair of HLA Class II genes,
consisting of one α and one β subunit;
wherein the cells are dendritic cells, macrophages or B-cells
Paragraph 100 of the specification recites
“the method comprises providing a population of cells that expresses either a single class I HLA allele, a single pair of class II HLA alleles, or a single class I HLA allele and a single pair of class II HLA allele” and “that other cell populations can be generated which are class I and/or class II deficient; and that “the population of cells are professional antigen presenting cells such as macrophages, B cells and dendritic cells”. Thus, the dendritic cells and B cells were from cell lines constructed with single pairs of MHC alleles are not dendritic cells and B cells from patient-derived cells. The other recitals of dendritic cells in the specification are for a dendritic cell vaccine.
35 USC § 112 rejections maintained
The rejections of claims 16-18, 32, 33, 39, 41, 42 and 45-48 and new claim 50 under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the written description requirement are maintained.
Applicant argue that regarding item (i) above, use of both cultured cells expressing a distinct single class I HLA allele and patient-derived cells in the claimed methods can be found throughout the specification as filed. Applicant argues that the specification clearly specifies that both populations of cells can be used in combination in Paragraph [0526] of the published U.S. application, which states "[t]he methodologies described herein provide a path toward addressing new questions relating to HLA ligandomes .... Applicants can also apply the workflows to enable the sequencing of HLA class I and class II peptides presented by patient-derived cell lines and primary tumor samples, which can provide an opportunity to make the observations more direct and personalized.
Applicant argues that the citation above uses the additive term "also," implying that the use of patient derived cells can be used in addition to single-allelic cell lines. Applicant argues that the application also explicitly categorizes processes disclosed in the application, including the claimed methods, as exemplary "workflows". Applicant argues that the abstract, states, "[t]hese streamlined experimental and analytic workflows enable direct identification and analysis of endogenously processed and presented antigens".
Applicant’s arguments have been considered but are not persuasive. The specification discloses that an overview of experimental workflow is provided in Fig. 1A-1D. Figure 1 involves a B cell line expressing single HLA class 1 allele. It is clear the focus of the specification involves isolating peptides from cell lines expressing a single HLA allele. Paragraph 526 does indicate that the workflow may also done to enable the sequencing of HLA class I and class II peptides presented by patient-derived cell lines and primary tumor samples. But it is a stretch to say that this paragraph supports the limitation “ training a machine with the HLA-allele specific binding neoantigenic peptide sequence database that consists of the peptides of from the HLA-peptide complexes of the first population of cultured cells expressing a distinct class I HLA allele and the peptides from the HLA-peptide complexes of the second population of patient derived cells. As Applicant has pointed out there is nowhere in the specification that specifically indicates training a machine with the HLA-allele specific binding neoantigenic peptide sequence database that consists of the peptides of from the HLA-peptide complexes of the first population of cultured cells expressing a distinct class I HLA allele and the peptides from the HLA-peptide complexes of the second population of patient derived cells. Given that the overflow disclosed in Fig. 1 involves the isolation of peptides and determining characteristics of the peptides, there does not even appear to be an inference that training a machine was done with both All of the data appears to be derived from cultured cells expressing a distinct class I HLA allele. Given that that Applicant appears to be arguing that the patentability of the present application depends from training a machine with peptides from both cultured cells expressing a distinct class I HLA allele and patient derived cells it would seem that the specification should at least specifically or inferentially indicate that the training would be done with peptides from both cultured cells expressing a distinct class I HLA allele and patient derived cells.
Furthermore, as described above, paragraphs 11-16 of the specification recite
(a) providing a population of cells expressing a single HLA allele;
(a) providing a population of cells expressing a single HLA class I allele;
(a) providing a population of cells expressing a pair of HLA Class II genes,
consisting of one α and one β subunit;
wherein the cells are dendritic cells, macrophages or B-cells
Paragraph 100 of the specification recites
“the method comprises providing a population of cells that expresses either a single class I HLA allele, a single pair of class II HLA alleles, or a single class I HLA allele and a single pair of class II HLA allele” and “that other cell populations can be generated which are class I and/or class II deficient; and that “the population of cells are professional antigen presenting cells such as macrophages, B cells and dendritic cells”. Thus the dendritic cells and B cells were from cell lines constructed with single pairs of MHC alleles are not dendritic cells and B cells from patient-derived cells. The other of recitals of dendritic cells in the specification are for a dendritic cell vaccine.
In response to Applicants argument that the abstract, states, "[t]hese streamlined experimental and analytic workflows enable direct identification and analysis of endogenously processed and presented antigens", the Abstract involved B cells expressing a single HLA class I allele. The isolated peptides from the B cells would have been endogenously processed and presented with the single HLA class I allele. It is not clear how the Abstract supports training a machine with peptides from both cultured cells expressing a distinct class I HLA allele and patient derived cells.
35 USC § 101 rejections maintained
The rejections of claims 16-18, 32, 33, 39, 41, 42 and 45-48 and 50 as not being directed to patent eligible subject matter under 35 USC § 101 are maintained
The claims recite “judicial exceptions” as a limiting element or step without reciting additional elements/steps that integrate the judicial exceptions into the claimed inventions such that the judicial exceptions are practically applied, and are sufficient to ensure that the claims amount to significantly more than the judicial exceptions themselves. In the instant case, the “judicial exception” include the abstract idea, “generating a prediction algorithm for identifying subject specific HLA binding neoantigenic peptides”, “generating an HLA-allele specific binding neoantigenic peptide sequence database comprising a plurality of sequences of peptides” and “training a machine with the HLA-allele specific binding neoantigenic peptide sequence database” which are not eligible for patent protection without significantly more recited in the claims.
A.
Applicant argues that the amended claims are patent eligible under Step 2B
Applicant argues that the amended claims contain a combination of active steps that represent additional elements that are not well-understood, routine, and conventional. The combination of steps recited in the amended claims therefore provide "specific limitation[s] other than what is well-understood, routine and conventional in the field." Applicant argues that the answer to Step 2B inquiries for determining whether claims are subject matter eligible under 35 U.S.C. § 101: is yes, and the amended claims are patent eligible.
B.
Applicant argues that Applicant submits that the Examiner has not met his burden of showing the claimed methods are "well-known, routine and conventional." Applicant argues that the standard for "well-known, routine and conventional" under Prong 2B is not the obviousness standard under 35 U.S.C. § 103. Applicant argues that as such, to show that a feature is well-known, routine and conventional for patent eligibility requires analysis beyond merely pointing to the 35 U.S.C. § 103 prior art.
Applicant states that during the interview, Specialist Riggs confirmed that the standard for well-understood, routine, and conventional is not the same standard as 35 U.S.C. § 103 obviousness. Specialist Riggs also pointed to MPEP 2106.07 (a)(IIl)(C), which outlines the evidentiary requirements in making a 35 U.S.C. § 101 rejection to show that a feature is well-understood, routine, and conventional. This passage highlights that any publication merely disclosing a claimed feature does not establish that said claimed feature is well-understood, routine, and conventional under Prong 2B. Applicant argues that the appropriate art to demonstrate a claimed feature is well-understood, routine, and conventional falls within a much narrower category - it includes secondary sources that describe the state of the art, not a combination of references including research papers.3 Specifically, MPEP 2106.07 (a)(Ill)(C) states:
A citation to a publication that demonstrates the well-understood, routine, conventional nature of the additional element(s). An appropriate publication could include a book, manual, review article, or other source that describes the state of the art and discusses what is well-known and in common use in the relevant industry. It does not include all items that might otherwise qualify as a "printed publication" as used in 35 U.S.C. 102. Whether something is disclosed in a document that is considered a "printed publication" under 35 U.S.C. 102 is a distinct inquiry from whether something is well-known, routine, conventional activity.
Applicant further argues that analysis under Prong 2B requires an analysis of the claim features as a whole and in combination and a showing that the claimed inventions as a whole is not well-known, routine and conventional. MPEP 2106.05(1), which discusses Prong 2B:
this approach considers all claim elements, the Supreme Court has noted that
"it is consistent with the general rule that patent claims 'must be considered as a whole."' Alice Corp., 573 U.S. at 218 n.3, 110 USPQ2d at 1981 (quoting Diamond v. Diehr, 450 U.S. 175, 188,209 USPQ 1, 8-9 (1981)). Consideration of the elements in combination is particularly important, because even if an additional element does not amount to significantly more on its own, it can still amount to significantly more when considered in combination with the other elements of the claim. See, e.g., Rapid Litig. Mgmt. v. CellzDirect, 827 F.3d 1042, 1051, 119 USPQ2d 1370, 1375 (Fed. Cir. 2016).
Applicant argues that the Examiner, in using disclosures from multiple references to support the rejection under 35 U.S.C. 101, has failed to meet his burden of showing the combination of active steps are well-known, routine and conventional. Applicant argues that when discussing the combination of claimed steps during the interview, Specialist Riggs again pointed the Examiner to the MPEP and reiterated that the Examiner must consider the steps of claim 16 as a whole and in combination.
Applicant argues that regarding item (i) above, use of both cultured cells expressing a distinct single class I HLA allele and patient-derived cells in the claimed methods can be found throughout the specification as filed. Applicant argues that the specification clearly specifies that both populations of cells can be used in combination in workflows disclosed in the application.
Applicant’s arguments have been considered but are not persuasive. The purpose for isolating the HLA-peptide complexes from the cultured cells expressing a distinct single class I HLA allele and isolating the HLA-peptide complexes from dendritic or B-cells of the patient-derived populations of cells are the same, to identify peptides which bind to particular HLA alleles to categorize which amino acids at which positions optimize binding of the peptide to the particular MHC allele. Determining which amino acids at which positions were relevant to binding to a specific HLA allele was well known in the art (Rovero et al Mol Immunol 31:549-554, 1994). One could not differentiate the amino acid structure of peptides from cultured cells expressing a single class I HLA allele and peptides from B cells of a patient. The advantage of using cell lines expressing a single class I HLA allele is that the peptides would be easier to isolate and identify than peptides from tumor cells or cell lines comprising more than one HLA allele. Furthermore, peptides could be identified that bound less common HLA alleles to determine which amino acids of a peptide are important for binding to the less common HLA alleles. Cell lines expressing a single class I HLA allele had been known for over 25 years from Applicant’s filing date (Shimizu et al, J Immunol 142:3320-3328, 1989).
Min disclose that as the instant methods are based on the analysis of sequences of known binders and non-binders, the predictive performance will continue to improve with accumulation of the experimentally verified binding/non binding peptides (paragraph 8). This ability to accommodate and scale with increasing amounts of data is critical for further refinement of the prediction ability of the method (Id). Thus, the ability to predict which peptides will bind to a particular MHC allele will increase with the accumulation of binding data. In this regard the immune epitope database (IEDB) contains more than 15,000 journal article and more than 704,000 experiments as of 2014 (abstract; Vita et al Nucleic Acid Research 43:D405-D412, 2014). Given that the IEDB would include the published and public amino acid sequences of peptides bound to particular HAL alleles the IEDB would necessarily include peptides from cultured cells expressing a single class I HLA allele and peptides from B cells of a patient. Given that the amino acid sequences listed in the present specification were published they were likely added to the IEDB. As previously discussed both cell lines, which would include cell lines expressing a single allele HLA and tumor samples have been used previously used to identify which amino acids are important for peptides which bind to the specific MHC allele.
The art discloses that peptides bound to isolated B cell tumor cells from patients (paragraphs 40, 64, 70 of Johnston; paragraphs 6, 39, 93, 145-147 of Yelensky).
The art discloses that peptides binding to a particular MHC can be identified using single-allele cell lines (paragraphs 35, 36, 93 of Yelensky; paragraphs 315, 316 of Bergeron; paragraph 476 of Fikes).
Schirle et al (Eur J Immunol 30:2216-2225, 2000) disclose that sources of peptides bound to HLA molecules are tumor cell lines and solid tumors (page 2217, 1st column). Schirle disclose transgenic mice expressing a single HLA allele page 2216, 2nd column to page 2217, 1st paragraph, 1st paragraph).
Both Yelensky and Johnston disclose peptides derived from that the tumor cells including B-cell lymphomas while Yelensky and Schirle discloses that both primary tumor cells and cell lines may be used to isolate and identify peptides that bind a specific HLA allele.
Thus, the sources for the isolation and identification of peptides bound to a particular HLA allele were known in the art as well as art disclosing both types of sources of isolated and identified peptides listed in the claims.
In addition, MPEP 2106.07 (a)(Ill)(D) states
A statement that the examiner is taking official notice of the well-understood, routine, conventional nature of the additional element(s). This option should be used only when examiners are certain, based upon their personal knowledge, that the additional element(s) represents well-understood, routine, conventional activity engaged in by those in the relevant art, in that the additional elements are widely prevalent or in common use in the relevant field, comparable to the types of activity or elements that are so well-known that they do not need to be described in detail in a patent application to satisfy 35 U.S.C. 112(a).
As discussed above, both cell lines, which would include cell lines expressing a single allele HLA and tumor samples have been used previously used to identify which amino acids of the peptides are important for binding to the specific MHC allele. The amino acid sequences of the peptides for a particular HLA from the different sources would likely overlap, given that it was known that particular amino acids at particular positions of a peptide are corelated with the binding of that peptide to a particular HLA allele. Thus, the amino acid sequences and the function of the identified peptides between the different sources of the peptides would overlap.
In addition, Applicant recites that peptides isolated using the method set out in Example 1 with peptides isolated using conventional techniques and present in the Immune Epitope Database (IEDB). The IEDB would include peptides isolated from cell lines and primary tumor cells. The cell lines would include cell lines expressing a single allele HLA. As previously discussed, cell lines expressing a single class I HLA allele had been known for over 25 years prior to Applicant’s filing date (Shimizu et al, J Immunol 142:3320-3328, 1989, cited previously).
35 USC § 103(a) rejections maintained
The rejection of claims 16-18, 32, 33, 39, 42 45, 47, 48 and 50 under 35 U.S.C. 103 as being unpatentable over Min et al US 2015/0278441 published October 1, 2015) and Yelensky et al (2017/0199961, published 13 July 2017, effective filing date 4 April 2016, cited previously) in view of Fikes et al (US 2004/0018971, published 29 January 2004) and Bergeron et al (US 2009/0028888, published 29 January 2009) are maintained.
The claims are drawn to a method of for generating a prediction algorithm for identifying HLA-allele specific binding peptides comprising training a machine with an HLA-allele specific binding peptide sequence database comprising sequences of peptides bound to an HLA of a population of cells, wherein each cell in the population of cells expresses a single class I HLA allele and wherein variables used to train the machine comprise the expression level of source proteins of the peptides within the population of cells.
Min teaches a system to predict peptide-histocompatability complex class (MHC) interaction uses high-order semi-Restricted Boltzmann Machines with deep learning extensions to efficiently predict peptide-MHC binding (paragraph 6). Min further disclose a method for peptide binding prediction includes receiving a peptide sequence descriptor and optional structural descriptor of MHC protein-peptide interaction; generating a model with one or an ensemble of high order neural networks; pre-training the model by high-order semi-Restricted Boltzmann machine (RBM) or high-order denoising autoencoder; and generating a prediction as a binary output or continuous output (paragraph 7). Min discloses that the input data is provided to a model layer which can be a linear model, a kernel SVM, or an ensemble of traditional feed-forward neural networks (paragraph 3). Min disclose that the methods allow integration of both qualitative (i.e., binding/non-binding/eluted) and quantitative (experimental measurements of binding affinity) peptide-MHC binding data to enlarge the set of reference peptides and enhance predictive ability of the method (paragraph 8). Min disclose that in order for the peptides to bind to a particular MHC, the sequences of the binding peptides should be approximately superimposable: contain similar amino-acids or strings of amino acids (k- mers) at approximately the same positions along the peptide chain (paragraph 30). Min further disclose that sequence of the descriptors corresponding to the peptide can be modeled as an attributed set of descriptors corresponding to different positions (or groups of positions) in the peptide and amino acids or strings of amino acids occupying these positions (paragraph 32). Min disclose that each position in a peptide is described by a feature vector, with features derived from the amino acid occupying this position/or from a set of amino acids (paragraph 34). Thus, Min teaches training a machine wherein the machine combines one or more linear models, support vector machines, decision trees and neural networks wherein the variables used to train the machine comprise amino acid frequency at specific positions of the peptide.
Yelensky teaches a presentation model that can comprise a statistical regression or a machine learning (e.g., deep learning) model trained on a set of reference data (also referred to as a training data set (paragraph 93). The training data set includes tissue-specific expression, expression of the TAP protein, ability of a peptide to bind the TAP protein, and stability of the peptide (paragraphs 329, 336-338, 360, Fig. 4). Yelensky discloses using databases to train a machine (paragraphs 158, 184, 483, 492)
One of ordinary skill in the art would have been motivated to apply Yelensky’s method for training a machine using variables such as the expression level of the source protein, ability of a peptide to bind the TAP protein and stability of the peptide to Min’s method for training a machine with an HLA-specific binding peptide sequence database of known binders and non-binders along with quantitative measurements of binding affinity because both Yelensky and Min concern training a machine with machine learning to predict binding of peptides to a specific class I HLA allele.
Bergeron discloses using cell lines that do not express endogenous HLA molecules transfected with an expression construct encoding a single HLA allele for obtaining the motif-bearing peptides correlated with the particular HLA molecule expressed on the cell (paragraphs 315-316). Bergeron discloses peptide anchor residues for binding to class I (paragraphs 233, 238).
Fikes disclose obtaining the motif-bearing peptides correlated with the particular HLA molecule expressed on the cell using cell lines that do not express any
endogenous HLA molecules transfected with an expression construct encoding a single HLA allele (paragraph 476). Fikes disclose peptide anchor residues for binding to an HLA molecule (paragraphs 48, 81, 85; Table I).
One of ordinary skill in the art would have been motivated to apply Bergeron and Fikes disclosure of using cell lines transfected with an expression construct encoding a single HLA allele to Min and Yelensky’s method for training a machine with machine learning to predict binding of peptides to a specific class I HLA allele because both Bergeron and Fikes disclose that using cell lines that do not express any endogenous HLA molecules transfected with an expression construct encoding a single HLA allele is an alternative method of isolating and identifying peptides bound to a particular HLA class I allele. It would have been prima facie obvious to combine Min and Yelensky’s method for training a machine with machine learning to predict binding of peptides to a specific class I HLA allele with Bergeron and Fikes’s use of cell lines that do not express any endogenous HLA molecules transfected with an expression construct encoding a single HLA allele to identify peptides bound to a specific HLA class I allele to have a method for generating a prediction algorithm for identifying HLA-allele specific binding peptides comprising training a machine with an HLA-allele specific binding peptide sequence database comprising sequences of peptides bound to an HLA of a population of cells, wherein each cell in the population of cells expresses a single class I HLA allele and wherein variables used to train the machine comprise the expression level of source proteins of the peptides within the population of cells.
The rejections of claims 16-18, 32, 33, 39, 41, 42 and 45-48 and 50 under 35 U.S.C. 103 as being unpatentable over Min et al US 2015/0278441 published October 1, 2015, cited previously) and Yelensky et al (2017/0199961, published 13 July 2017, effective filing date 4 April 2016, cited previously) in view of Bergeron et al (US 2009/0028888, published 29 January 2009, cited previously) and Fikes et al (US 2004/0018971, published 29 January 2004, cited previously) in further view of Johnston (US 2015/0079119, published 19 March 2015) and Rammensee et al (US 2017/0022251, published 26 January 2017, effective filing date 25 June 2015) are maintained.
Neither Min, Yelensky, Fikes nor Bergeron disclose the mutant HLA allele HLA A*03:01 nor that the sequencing is performed by LC-MS/MS.
Johnston disclose using LC-MS/MS to identify peptide bound to HLA alleles (paragraphs 54,233, 234, 239)
Rammensee disclose using LC-MS/MS to identify peptide bound to HLA alleles including HLA A*03:01 (144-148, 359-364; Table 5B).
One of ordinary skill in the art would have been motivated to apply Johnston and Rammensee’s method of detecting peptides bound to HLA alleles using LC-MS/MS to Min, Yelensky, Fikes and Bergeron’s method for generating a prediction algorithm for identifying HLA-allele specific binding peptides comprising training a machine with an HLA-allele specific binding peptide sequence database because Min, Yelensky, Fikes and Bergeron, Johnston and Rammensee teach isolating and identifying peptides bound to class I HLA alleles. It would have been prima facie obvious to combine Min, Yelensky, Fikes and Bergeron’s method for generating a prediction algorithm for identifying HLA-allele specific binding peptides comprising training a machine with an HLA-allele specific binding peptide sequence database with Johnston and Rammensee’s method of detecting peptides bound to HLA alleles using LC-MS/MS to have a method for generating a prediction algorithm for identifying HLA-allele specific binding peptides comprising training a machine with an HLA-allele specific binding peptide sequence database comprising sequences of peptides bound to an HLA of a population of cells using LC-MS/MS, wherein each cell in the population of cells expresses a single class I HLA allele and wherein variables used to train the machine comprise the expression level of source proteins of the peptides within the population of cells.
Applicant argues that neither Min, Yelensky, Fikes, Bergeron, Johnston nor Rammensee, either alone or in combination, teach or suggest obtaining multiple populations of cells, wherein the multiple populations of cells comprise a first population of cultured cells expressing a distinct single class I HLA allele comprising HLA-peptide complexes and a second population of patient-derived cells comprising HLA-peptide complexes, wherein the peptides of the HLA-peptide complexes are endogenous peptides, as recited in independent claim 16. Applicant argues that none of the references, either alone or in combination, teach or suggest isolating the HLA-peptide complexes from the cultured cells expressing a distinct single class I HLA allele and isolating the HLA-peptide complexes from dendritic or B-cells of the patient-derived cells; (iii) isolating the endogenous peptides from the HLA-peptide complexes isolated in (ii) and sequencing the endogenous peptides by mass spectrometry.
Applicant argues that Bergeron discloses using cell lines that do not express endogenous peptides from cells encoding a single HLA allele for obtaining the motif-bearing peptides correlated with the particular HLA molecule expressed on the cell. Applicant also argues that in paragraph 476 of Fikes, the peptide of interest in the HLA-peptide complexes from the cells that express only a single type of HLA molecule is a known, exogenous, and overexpressed antigen, not an endogenous peptide according to the instantly claimed method. Applicant argues that according to paragraphs 474-476 of Fikes, the cells that express only a single type of HLA molecule are "infected with a pathogenic organism or transfected with nucleic acids that express the tumor antigen of interest" or cell lines that do not express endogenous HLA molecules can be transfected with an expression construct encoding a single HLA allele" and then "infected with a pathogenic organism or transfected with nucleic acid encoding an antigen of interest to isolate peptides corresponding to the pathogen or antigen of interest that have been presented on the cell surface."
Applicant’s argument has been considered but is not persuasive. The extended phrase given by Applicant concerning Fikes was “they may be infected with a pathogenic organism or transfected with nucleic acid encoding an antigen of interest to isolate peptides corresponding to the pathogen or antigen of interest that have been presented on the cell surface. (paragraph 476). Fikes does not disclose that they must be infected with a pathogenic organism. A prior art reference is relevant for all its teachings, not only its examples. Merck & Co. v. Biocraft Labs., Inc., 874 F.2d 804, 807 (Fed. Cir. 1989) (holding that both preferred and unpreferred embodiments must be considered). Fikes discloses a system for identifying peptides that bound to a specific HLA allele. In addition, the cell line comprising cells that express only a single type of HLA molecule would necessarily include endogenous peptides, which would likely be the majority of the peptides bound to the HLA allele. The fact that Fikes was primarily interested in determining whether their peptides of interest bound to a specific HLA allele does not negate the fact that the cell line could not be used to identify other peptides from endogenous proteins. Fikes disclosed using the cell line for his particular purpose. However, the advantages of using a cell line comprising cells that express only a single type of HLA molecule would be identical for the purposes of identifying the anchor residues and other relevant amino acids for use in prediction algorithms to identify which peptides would likely bind to a particular HLA allele.
In response to Applicant’s argument that Bergeron discloses using cell lines that do not express endogenous peptides from cells encoding a single HLA allele for obtaining the motif-bearing peptides correlated with the particular HLA molecule expressed on the cell, the cell lines encoding a single HLA allele would necessarily include endogenous peptides presented in the context of the single HLA allele. In fact, the majority of the peptides presented in the context of the particular HLA allele would likely be endogenous. As described above for Fikes, Begeron discloses that the particular cell lines can be used to isolate and identify MAGE peptides. However, as described above, the advantages of using a cell line comprising cells that express only a single type of HLA molecule would be identical for the purpose of identifying the anchor residues and other relevant amino acids for use in prediction algorithms to identify which peptides would likely bind to a particular HLA allele. The fact that the cell lines encoding a single HLA allele can be used to identify HLA allele specific peptides from a particular protein does not detract from their use in identifying the anchor residues and other relevant amino acids for use in prediction algorithms to identify which peptides would likely bind to a particular HLA allele.
In addition, Applicant argues that none of the references, either alone or in combination, teach or suggest isolating endogenous peptides from HLA-peptide complexes isolated from cultured cells expressing a single class I HLA allele. Applicant argues that based on the combined teachings of the cited references, a skilled artisan would have had no expectation of success that the single class I HLA allele cells that overexpress a known exogenous antigen could be used in a method for generating a prediction algorithm for identifying subject specific HLA binding neoantigenic peptides in which a variable used to train the machine comprises the expression level of endogenous source proteins. Applicant argues that a skilled artisan would not have been able to determine the expression level of the endogenous source proteins within the populations of cells containing the endogenous peptides of the HLA-peptide complexes at least because the source protein is overexpressed in the cells expressing a single class I HLA allele described in Bergeron and Fikes. Applicant argues that a skilled artisan would have had no motivation to combine the cells of Bergeron and Fikes expressing a single class I HLA allele that overexpress an exogenous antigen with the teachings of the other cited references to arrive at a method of generating a prediction algorithm for identifying subject specific HLA binding neoantigenic peptides that requires training the machine with the expression level of endogenous source proteins as recited in independent claim 16.
Applicant also argues that the amended claims also take advantage of the incorporation of data from both cell lines expressing a distinct single class I HLA allele comprising HLA-peptide complexes and patient-derived (multi-allelic) cells comprising HLA-peptide complexes. Applicant argues that data from cell lines expressing single HLAs does not require deconvolution to determine which HLA the MS-observed peptide was bound, which has advantages over data from patient derived cells expressing multiple HLAs in which deconvolution is typically needed to predict which of these multiple HLAs the MS-observed peptide was bound. However, on the other hand, data from cell lines expressing single HLA will likely have a different expression profile from that of a patient cancer cell, also potentially producing bias.
Applicant argues that neoantigens can be used therapeutically in immunogenic compositions to direct the subject's immune response to specifically attack tumor tissue. When administered to a patient, the neoantigens bind to the HLA alleles present in the patient where they are presented to T cells. The T cells which recognize the neoantigens are then primed to attack tumor tissue which also present the neoantigens. Efficiently choosing which neoantigens to administer as part of an immunogenic
composition requires the ability to predict which neoantigens bind to the HLA alleles present in a patient.
Applicant argues that prediction algorithms have been used to try and determine which neoantigens are likely to bind to HLA alleles present in a patient. Applicant argues that several factors limit the power of prediction algorithms to identify which neoantigens will be presented by HLA molecules. Applicant argues that the factors include the following:
• First, the provenance of peptide data upon which these algorithms are trained is diverse. The algorithms most commonly used at the priority date of the instant application were trained almost exclusively on measurements of biochemical affinity of synthetic peptides. Applicant argues that in-silica methods do not provide information on peptide parameters, such as protein expression, that are affected by the disease context. Applicant argues that the inclusion of such information from patient derived cells can provide a more accurate details of protein parameters in disease.
• Second, many existing prediction algorithms have focused on predicting binding but may not fully take into account endogenous processes that generate and transport peptides prior to binding.
• Third, the number of binding peptides for many HLA alleles is too small to develop a
reliable predictor. As an example, see Bergeron, cited by the Examiner, which utilizes
peptides derived from MAGE-A9, rather than a diverse set of peptides.
Applicant found that training a machine with a database comprising peptide sequences
obtained from cell lines expressing multiple alleles as well as those expressing a single class I HLA allele, wherein variables used to train the machine comprise peptide sequence and source protein expression of the peptide, provides improved prediction algorithms and methods for identifying neoantigen-comprising peptides.
Applicant’s arguments have been considered but are not persuasive. As an initial matter, Yelensky discloses prediction algorithms for identifying cancer and patient specific peptides which would likely bind to the patient’s HLA allele with an affinity that may induce an immune response to those cancer and patient specific peptides. It is noted that a prediction algorithm for identifying HLA allele specific binding subject specific HLA binding neoantigenic peptides would also function as a prediction algorithm for identifying specific HLA allele specific peptides that would not be subject specific. The important factors determining which peptides bound would be the amino acid sequence of the cellular processed peptides. As discussed above, the anchor residues and related amino acids and their positions would determine whether the particular peptide would bind the specific HAL allele with sufficient affinity. To be subject-specific the algorithm would have to have to be able to differentiate cancer-specific peptides from wild type peptides. The algorithm would further have to ensure that the mutated amino acid was not an anchor residue which would likely preclude the peptide from being recognized by the patient’s T cells. The present claims do not appear to have such differentiating steps. Furthermore, as previously discussed, the function of Yelensky’s algorithm was to identify peptides which were cancer and patient-specific.
It is noted that the elected invention was drawn to a method for generating a prediction algorithm for identifying HLA allele specific binding peptides while the unelected invention was for a method of identifying from a given set of neo-antigens comprising peptides the most suitable peptides for preparing an immunogenic composition for a subject.
In response to Applicant’s arguments that previous prediction algorithms did not take into account protein expression, and endogenous processes that generate and transport peptides prior to binding are not taken into account, as discussed above, both Min and Yelensky disclose such factors when generating their prediction algorithms. Yelensky disclose that their training data set includes tissue-specific expression, expression of the TAP protein, ability of a peptide to bind the TAP protein, and stability of the peptide.
In response to Applicant’s argument that the amended claims also take advantage of the incorporation of data from both cell lines expressing a distinct single class I HLA allele comprising HLA-peptide complexes and patient-derived (multi-allelic) cells comprising HLA-peptide complexes, as discussed above the IEDB would include peptides from both cell lines expressing a distinct single class I HLA allele and patient-derived (multi-allelic) cells.
In response to Applicant’s argument that a skilled artisan would have had no expectation of success that the single class I HLA allele cells that overexpress a known exogenous antigen could be used in a method for generating a prediction algorithm for identifying subject specific HLA binding neoantigenic peptides in which a variable used to train the machine comprises the expression level of endogenous source proteins, it is not clear why one of skill in the art using mass spectrometry would not have had a reasonable expectation of success in isolating and identifying endogenous peptides from single class I HLA allele cells. First, the methods would be simpler than isolating and identifying endogenous peptides from cell lines expressing several HLA alleles. There would no need to isolate individual HLA molecules or use deconvolution to identify specific HLA peptides bound to a particular HLA molecule. Identifying the peptides by mass spectrometry would then be It is likely that a skilled artisan would have a greater expectation of success using single class I HLA allele cells to determine which peptides are bound to a particular HLA molecule.
In response to Applicant’s argument that data from single allelic cell lines is not present in the first two provisional applications of Yelensky, as stated above, the priority date for the present claims is October 18, 2018. Furthermore, both Fikes and Bergeron disclose single HLA allelic cell lines.
In response to Applicant’s argument that data from cell lines expressing single HLA will likely have a different expression profile from that of a patient cancer cell, also potentially producing bias, it is noted that most cell lines are derived from cancer cells. Thus, the peptides isolated and identified from a cell line expressing a particular HLA allele would likely overlap. Applicant has not provided any evidence that there would be no overlap between the peptides isolated and identified from cell lines expressing a particular HLA allele and a cell from a primary tumor expressing that particular HLA allele. As discussed above, the prediction algorithm is based on the fact that particular amino acids at particular positions in a sequence determine whether that peptide will bind to a specific HLA allele. It’s irrelevant whether that particular peptide is from a cell line or a peptide from a cell from a primary tumor.
In response to Applicant’s argument that data from cell lines expressing single HLAs does not require deconvolution to determine which HLA the MS-observed peptide was bound, which has advantages over data from patient derived cells expressing multiple HLAs in which deconvolution is typically needed to predict which of these multiple HLAs the MS-observed peptide was bound, the use of cell lines expressing single HLAs for determining which peptides bound to a particular HLA molecule was well known prior to Applicant’s filing date. These same advantages would be applicable for transfecting exogenous antigens into cell lines expressing single HLA. It would be easier to isolate and identify peptides which bound to a particular HLA molecule. Identification of the peptides may be easier given that the structure of the exogenous antigen was known, but with mass spectrometry identification of unknown peptides bound to a particular HLA molecule in cell lines expressing single HLA is relatively straightforward. As previously discussed above, cell lines expressing a single class I HLA allele had been known for over 25 years from Applicant’s filing date (Shimizu et al, J Immunol 142:3320-3328, 1989, cited previously)
Applicant also argues that each of Fikes, Bergeron, Johnston, Sette and Rammensee are focused on therapeutics to shared antigens, such as tumor associated antigens, universal to unrelated individuals. Applicant argues that these references are contrary to a method of generating an algorithm that can identify subject specific HLA binding neoantigenic peptides for use in personal cancer approaches that rely on endogenous cancer specific mutation(s). Thus, one of ordinary skill in the art combining these references, even in combination with Min and Yelensky, would not have
arrived at the methods of claim 16.
In response neither Fikes, Bergeron, Johnston nor Sette teach away from Yelensky’s method for identifying cancer and patient specific peptides that bound the same HLA allele as expressed on the cancer patient.
Applicant contends that the working Examples highlight the claimed method and unexpected results of the amended claims
Applicant points the Examiner to Example 1 of the application, describing methods of generating peptide sequence databases, each database comprising peptide sequences that all unambiguously bind to a specific HLA molecule. Thus, paragraph [00432] of Example 1 explains that the Applicant immunoaffinity-purified and sequenced HLA-associated peptides from class I deficient cells that were stably transduced to express a single class I HLA allele. Importantly, the inventors' use of cells expressing a single class I HLA allele allows peptides that do not closely match known motifs to be confidently reported as binders to that class I HLA molecule. Paragraph [00454] of the present application further explains the significance of this difference as "Although LC-MS/MS-based approaches to identify the HLA-peptidome have long been employed, these studies have typically utilized primary cells or cell lines expressing the full complement of HLA molecules, making it challenging to distinguish allele-specific characteristics related to peptide display." Applicant argues that the steps used in Example 1 are recited in steps (a)-(d) of claim 16, therefore, the amended claims are commensurate in scope with the results described in the Examples. Examples 2-4 provide a comparison of peptides isolated using the method set out in Example 1 with peptides isolated using conventional techniques and present in the Immune Epitope
Database (IEDB). In particular, paragraph [00442], explains that "the LC-MS/MS data captures new peptide-binding motifs not reflected in the IEDB". Furthermore, paragraph [00443] explains that highly expressed proteins are more likely to be processed and presented by the HLA class I pathway. Thus, Examples 2-4 identify peptide sequence and source protein expression as variables that determine the likelihood of a peptide to bind to a specific HLA class I molecule. These variables are recited in claims 1 and 6. Applicant argues that based on these findings, Applicant developed prediction algorithms for predicting whether peptides bind to class I HLA alleles. Example 5 describes the development of two new prediction algorithms, MSintrinsic and MSintrinsicEC. MSintrinsicEC was generated according to the method of claim 1. Paragraph [00453] explains that both MSintrinsic and MSintrinsicEC outperformed the
prior prediction algorithms NetMHC-4.0 and NetMHCpan-2.8, with an average positive predictive value (PPV) improvement of 20 and 30 percentage points for 'MSintrinsic' and 'MSintrinsicEC', respectively, in an internal 5-fold cross validation with 999n decoys
Applicant’s arguments have been considered but are not persuasive. As an initial matter, as disclosed in the art, isolating and identifying HLA-associated peptides from class I deficient cells that were stably transduced to express a single class I HLA allele was well known in the art. As discussed previously, both Bergeron and Fikes disclose using cell lines transfected with an expression construct encoding a single HLA allele to identify peptides that bind a particular HLA allele. Yelensky disclosure of using pan-HLA antibodies and deconvolution from multiallelic cells does not detract from the wide spread use of using cell lines transfected with an expression construct encoding a single HLA allele to identify peptides that bind a particular HLA allele. Furthermore, as disclosed in the art, the variables used to train the machine comprise the expression level of source proteins, the sequence of the peptides, and cleavability of the peptides within the population of cells, were known in the art. That highly expressed proteins are more likely to be processed and presented by the HLA class I pathway was already known in the art. The IEDB database was constructed using data from peptides isolated from specific HLA molecules, more than likely with some of the data obtained from the use of cells expressing a single class I HLA allele. However, a database is only as good as the data used to construct the algorithm used to identify likely peptides that would bind a particular class I HLA allele. The more data obtained, the better the algorithm would be at identifying peptides likely to bind a particular class I HLA allele. With more data on which peptides bound to a particular MHC allele, a better pattern would develop as to which amino acids at which position would likely bind the particular MHC allele. It is also likely that the type of cell would influence the population of peptides bound to the particular MHLA allele. It is noted that the claims do not list specific peptide-binding motifs that are not reflected in the IEDB and would encompass peptides isolated and identified using single-HLA allele cell lines that are listed in the IEDB. If Applicant’s invention is the identification of peptides bound to unique HLA alleles that have not been previously examined the claims should be amended to reflect the new peptide-binding motifs. As previously discussed above, cell lines expressing a single class I HLA allele had been known for over 25 years prior to Applicant’s filing date (Shimizu et al, J Immunol 142:3320-3328, 1989, cited previously).
Furthermore, an algorithm will never be as accurate as actual experiments to identify peptides bound to a particular class I HLA allele, whether it was using isolated HLA alleles from multi-allelic cells or from cells expressing a single class I HLA allele. Thus, more data on the peptides bound to a particular class I HLA allele would be expected to increase the predictability of a particular algorithm for identifying peptides likely to bind a particular HLA allele. And actual experimental results to identify peptides bound to a particular class I HLA allele will be more accurate at identifying peptides bound to a particular class I HLA allele than a prediction algorithm.
NEW REJECTIONS:
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
Claims 16-18, 32, 33, 39, 41, 42 and 45-48 and 50 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Section (c) recites (v) training a machine with the HLA-allele specific binding neoantigenic peptide sequence database. However section (iv) recites generating an HLA-allele specific binding neoantigenic peptide sequence database comprising a plurality of sequences of peptides, wherein the wherein the plurality of sequences of peptides comprise the peptides of (iii) from the HLA-peptide complexes of the first population of cultured cells and the peptides of (iii) from the HLA-peptide complexes of the second population of patient derived cells. However, it appears as if the second populations comprises peptides that could be derived from both class I and class II HLA alleles. It’s not clear if generating an HLA-allele specific binding neoantigenic peptide sequence database involves only peptides presented by MHC class I alleles or peptides presented by both MHC class I and MHC class II alleles and whether training a machine with the HLA-specific binding peptides involves training a machine with peptides presented by both MHC class I and MHC class II alleles.
In addition, the preamble recites a method for generating a prediction algorithm for identifying subject specific HLA binding neoantigenic peptides. However, there are no method steps which could be used to identify subject specific HLA binding neoantigenic peptides from peptides which are not subject specific HLA binding neoantigenic peptides. It is likely that the majority of peptides identified by the claimed method would be peptides which are not subject specific HLA binding neoantigenic peptides.
Summary
Claims 16-18, 32, 33, 39, 41, 42 and 45-48 and 50 stand rejected
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Mark Halvorson whose telephone number is (571) 272-6539. The examiner can normally be reached on Monday through Friday from 9:00 am to 6:00 pm. If attempts to reach the examiner by telephone are unsuccessful, the examiner's supervisor, Gregory Emch, can be reached at (571) 270-3503. The fax phone number for this Art Unit is (571) 272-8149.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free).
If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/MARK HALVORSON/Primary Examiner, Art Unit 1646