Section 3(j) of the Patents Act, 1970 bars patents on animals and any part thereof, and a cell grown in a dish is the hardest biotech case it produces. Harvard’s cell claim was refused on that clause, and the Delhi High Court sent it back in 2026 undecided.
That remand came three and a half years after the refusal, and it sends the case back to claims that were already on the file in August 2021. The applicant had filed them with its post-hearing written submissions, and the refusal recorded them as an alternative set before reasoning only on the claims they replaced. Harvard is therefore before the Controller again, nine years after entering India, with the eligibility question still open.
That eligibility question is governed by Section 3(j), one of the categories of non-patentable subject matter in Section 3. The clause carries the option India took under Article 27.3(b) of the TRIPS Agreement, which permits a member to exclude plants and animals while protecting micro-organisms. It works with Section 3(c), which excludes the discovery of any living thing occurring in nature, and the two together leave one narrow opening. A micro-organism qualifies, and only a modified one, because an isolated microbe is a discovery rather than an invention.
Where a cell sits in that opening is unsettled, and the Harvard file fell into the gap. A pancreatic β cell inside a human body is plainly part of an animal. The Controller says that growing the same cell from a stem cell in a dish changes the setting and not the thing. The applicant says that a cell with a gene expression profile no natural β cell has never existed in an animal at all.
Both positions survived the appeal intact, because the Delhi High Court disposed of it on the Controller’s failure to consider the amended claims and recorded that it had not gone into the merits. The Controller General published draft biotechnology examination guidelines in 2026, and neither of the two illustrations under Section 3(j) in them concerns a cell.
The Controller’s grounds for refusing Harvard’s biotech patent application
The Controller refused the application under four provisions at once, and only one of them was about biology. The order of 25 August 2022 held that Patent Application No. 201617000758, titled “SC-β Cells and Compositions and Methods for Generating the Same”, failed Sections 3(j), 3(e), 10(4) and 10(5) of the Patents Act, 1970, and refused it under Section 15. Section 3(j) was the eligibility objection, and the other three went to how the invention had been claimed and described. Objections stacked in this way are ordinary in Indian biotechnology patent prosecution.
The file by then was nine years old. Harvard had claimed priority from a United States application of June 2013, filed the international application PCT/US2014/041992 in June 2014, and entered India in January 2016. The first examination report issued in February 2020, and the reply with amended claims went in that August. The hearing took place in August 2021, written submissions carrying a further amended set followed within the fortnight, and the refusal came a year after that.
Whether a cell grown in a dish is a part of an animal is not an abstract question for anyone prosecuting a cell claim in India. Take an Indian cell-therapy company that files a claim to an engineered immune cell grown from a donor’s blood through a six-step protocol, reciting the surface markers and the secretion profile that no cell in the donor’s body carries. The first examination report objects under Section 3(j), on the ground that the product is a part of a human being.
The company answers that no cell of that description exists in a human body, and files the assay data showing it. The examiner replies that the starting material came out of a person, and that the protocol pushed those cells along a route they were already capable of taking.
Two difficulties then appear on the record, and they are separate. One is the eligibility question the objection raises. The other is that the company defined its product against what the donor’s body makes without stating that baseline anywhere, which leaves the claim open under Section 10(4) whichever way eligibility goes. On the first the company has an arguable answer, and on the second it has none until it amends.
Composition claims attract Section 3(e) as well as Section 3(j)
A composition claim attracts Section 3(e) on top of Section 3(j), because that clause asks whether the substance does more than aggregate the properties of its components. Harvard’s original claim 1 was a composition comprising a non-native pancreatic β cell together with one or more pharmaceutically acceptable carriers, additives or diluents, in the form of an in vitro cell cluster. The Controller found that the complete specification disclosed neither a composition as such, nor the components, nor their amounts, nor any data showing a combined effect beyond the aggregation of their properties.
Section 3(e) reaches a claim of that shape whether or not the ingredients are known. The Madras High Court read the clause that way in Novozymes v. Assistant Controller of Patents and Designs in September 2023, noting that the adjective “known” is used as a qualifier in Sections 3(d), 3(f) and 3(p) and is conspicuous by its absence from Section 3(e). The object of the clause, on that reading, is to exclude from patent eligibility any composition claim for a substance that merely exhibits the aggregate properties of its constituents, and the draft biotechnology guidelines of 2026 reproduce both passages.
The same wrapper carried the Section 10(4) and 10(5) findings, because the Controller held the technical disclosure missing in the claimed composition. The carrier language therefore attracted three of the four objections without answering the eligibility one. Harvard’s answer, filed after the hearing, was to delete it, and that deletion is what the appeal eventually turned on.
Delhi High Court remanded Harvard v Controller without ruling on Section 3(j)
The Delhi High Court set the refusal aside because the Controller had decided the case on claims the applicant had already replaced, and it decided nothing about whether a laboratory-grown β cell is patentable. Justice Tejas Karia delivered the judgment on 28 March 2026 in President and Fellows of Harvard College v. Controller General of Patents, Designs and Trade Marks, C.A.(COMM.IPD-PAT) 493/2022, an appeal under Section 117 of the Patents Act, 1970. The judgment records at paragraph 13 that the refusal had mentioned the amended claims as an “Alternative set of Claims” and had not considered them.
What the amendment did is set out at paragraph 14, where the two versions of claim 1 sit side by side. The original claimed a composition comprising the cell plus carriers, additives or diluents, in the form of an in vitro cell cluster. The amended claim drops all of that and claims the non-native pancreatic β cell itself, defined by its crystalline insulin granules, its expression of the genes INS, PDX1, NKX6-1 and ZNT8, its in vitro glucose stimulated insulin secretion response to a first glucose challenge, and its failure to express somatostatin or glucagon.
At paragraph 16 the Court took that change to its consequence. The refusal had reasoned under Sections 3(j) and 3(e) that the specification disclosed neither a composition per se, nor all the components, nor their amounts, nor data showing a combined effect beyond the aggregation of their properties. If the amended claims were considered, the Court held, “the entire basis of the Impugned Order may undergo change”. The direction at paragraph 19 set the order aside, remanded the file, and required a detailed order within six months of communication of the judgment after giving the applicant a hearing.
Paragraph 20 then closed off the reading most reports have given this case. The Court clarified that it had not considered the merits of the respective cases, and that the application was to be decided in accordance with law “without being influenced by any observations made in this Judgement”. Nothing in the judgment holds that a stem cell derived β cell escapes Section 3(j), and nothing in it holds the opposite either.
Unconsidered amended claims are why the Section 3(j) refusal was set aside
The refusal fell because the Controller never engaged with the amended claims, and two earlier decisions supplied the rule the Court applied to that failure. In Jitendra Kohli v. The Controller of Patents, neutral citation 2022:DHC:1904, the Delhi High Court found that the Assistant Controller had not taken the amended claims into consideration when deciding the fate of the application, held that this “is clearly a glaring error”, set the order aside and directed a fresh order after duly considering the amended claims on all counts including novelty, inventive step and patentability.
The same failure drew a stronger direction from the Madras High Court in Akebia Therapeutics INC v. The Controller of Patents and Designs, CMA(PT)/64/2024, decided on 20 March 2025, where the amendment had changed the claims from a method of treating anaemia to a pharmaceutical composition for treating it. The Court held that a rejection resting substantially or solely on the fact of such an amendment is not tenable, and that a remand was necessary because objections on inventive step and under Sections 3(i) and 3(e) had not been determined. It directed that an officer other than the one who issued the impugned order carry out the reconsideration, in order to preclude the possibility of predetermination, and recorded for the avoidance of doubt that no observation had been made on the merits.
The practical rule for anyone prosecuting an application is that an amendment changing the nature of the claim resets the examination, and a Controller who reasons past it hands the applicant an appeal. How far the applicant can take that amendment is set by the limits on amending a patent specification in India.
Examiners can still treat an engineered cell as an animal part under Section 3(j)
The Controller’s case against this cell claim survives the remand untouched, because the appeal turned on procedure and left every substantive objection standing. The core of it, recorded at paragraph 7.3 of the judgment, is that Section 3(j) excludes plants and animals in whole or any part thereof “irrespective of whether those parts are cultured, maintained, or differentiated in vitro”. Accepting that such cells are synthetic, the Controller argued, “would create a loophole whereby any naturally derived biological material could be rebranded as synthetic simply by laboratory handling”.
The argument has a second limb at paragraph 7.10, and it concedes more than a weak case would. In vitro differentiation does rely on specialised culture media and laboratory conditions, and the Controller accepts that these are not a direct replication of what occurs within the body. The answer is that the ultimate product is biologically identical to the β cell that naturally arises in vivo, so the laboratory process “merely triggers a natural developmental pathway inherent in stem cells” and does not give rise to a novel product unknown to nature.
Where that leaves the micro-organism carve-out is addressed at paragraph 7.4. Microbes are single-celled organisms, an animal is a multicellular eukaryote, and a genetically modified human cell cannot be equated with a genetically modified microbe because it falls on the far side of that line. The point has force in India precisely because the carve-out was read generously here first. The Calcutta High Court held in 2002, in Dimminaco A.G. v. Controller of Patents and Designs, that a process for preparing a vaccine was patentable although the end product contained a live organism, and that decision opened the microbiological route in India.
The definiteness objection that rides with Section 3(j)
A second objection travels with the eligibility one, and it does not disappear if the applicant wins on Section 3(j). The Controller’s submissions at paragraphs 7.6 to 7.8 attack the word “non-native” itself. The specification defines a non-native pancreatic β cell against a native one, and it does not define “native”.
Four points are made about that gap. The specification sets no standard for “native” or for a gene expression profile. There is no standard gene expression profile for a native cell in any event, so the comparator is itself a variable. Testing a cell for the non-native feature against every native variation would be an undue burden.
The fourth point uses the applicant’s own document. The Controller noted that the summary of the invention in the complete specification states that the non-native pancreatic β cell “is not genetically modified”, which on the Controller’s reading means the claim can read on a natural cell isolated from a human being. Nothing in the specification identifies the specific cell type, the phase at which the claimed genes and insulin crystals are present, or the expression level needed to obtain the composition.
None of this is a biology argument, which is why it matters to anyone drafting a cell claim. It is an objection under Section 10(4), which requires the claims to be definitive, and it would survive a ruling that engineered cells are eligible subject matter. The Controller’s position on this file is that the claim is unclear whichever way eligibility goes.
Myriad read against the biotech applicant
The Controller borrows the leading American authority on this question and turns it the other way. In Association for Molecular Pathology v. Myriad Genetics Inc., the United States Supreme Court concluded in 2013 that a claimed invention for the discovery of the precise location and sequence of two human genes was patent-ineligible, while synthetically created complementary DNA was eligible. The submission at paragraph 7.3 is that the distinction drawn there was between isolating or differentiating natural material and producing a genuinely synthetic construct, and that the SC-β cells, however cultured, remain natural derivatives. On that reading nothing in the Harvard file was synthesised at all, because a natural developmental pathway was run in a dish.
The force of the point is that it accepts the applicant’s own framework and then denies the applicant the benefit of it. Both sides agree that a man-made construct is eligible and that an isolated natural product is not, so the dispute reduces to which side of that line a stem cell derived β cell falls on. The Controller says the answer is fixed by what the product is, and the applicant says it is fixed by what was done to make it. Neither proposition was ruled on, and the Controller is free to press the first again on remand.
Applicants answer Section 3(j) by showing the construct does not occur in nature
The applicant’s answer is that Section 3(j) reaches natural parts of animals and essentially biological processes, and that a cell which exists nowhere in nature is neither of those things. That reading takes the clause on its own three limbs, at paragraph 8.3 of the rejoinder. The claimed SC-β cells are not whole animals, they are not natural parts of one, and they are not generated by essentially biological processes; they are the product of a multi-step laboratory-directed differentiation process.
Paragraph 8.5 answers the natural-pathway argument directly. Reprogramming and differentiating stem cells into β cells is an artificial, multi-stage laboratory protocol involving growth factors, culture conditions and gene regulation, which on the applicant’s case is neither essentially biological nor a natural developmental event. The mere use of a natural starting material does not bar patentability of a novel, non-naturally occurring product obtained from it. Paragraph 8.10 then puts the conclusion in the words that have since been attributed to the Court: the invention “represents a man-made construct (non-native), not an essentially biological process or natural material”, so the claims fall outside Section 3(j) and are patent-eligible.
BTS Research and the totality of human intervention under Section 3(j)
One Indian judgment has decided this question on the merits rather than remanding it on procedure, and it is the authority the applicant leaned on hardest. The Calcutta High Court decided BTS Research International Pty Ltd v. The Controller General of Patents, Designs and Trade Marks, IPDPTA 56 of 2023, on 3 April 2025. The subject was a method of generating tri-hybrid cells by fusing three somatic cells, at least two of them of different types, drawn from human and mouse sources.
The Court’s test for “essentially biological” is the part of that decision which has outlived it. A process is essentially biological if it consists entirely of natural phenomena such as crossing or the like, and the question of how much human intervention takes a process out of that category “can only be judged on the basis of the essence of the invention after taking into account the totality of human intervention and its impact on the result achieved”. Each step of an invention might be characterised as biological, and the arrangement of steps as a whole must still represent an essential modification of a known biological process. An invention has to be judged as a whole.
On the product side the Court rejected the “part of an animal” characterisation, holding that the impugned order had failed to address the artificial nature of the process and had ignored that tri-hybrid cells, being artificially produced and not naturally occurring, do not fall within the meaning of plants or animals. The refusal was set aside and remanded for reconsideration within twelve weeks. That outcome was itself a remand, and the Calcutta High Court did not direct a grant.
The Controller’s answer to BTS Research sits at paragraph 7.10 of the Harvard judgment, and it is a distinction rather than a denial. Tri-hybrid chimeric cells fused from human and mouse somatic cells are a construct that does not and cannot exist in nature, which is a level of artificial genetic manipulation that takes them outside Section 3(j), whereas a β cell can exist in nature. Drawing the line at the organism rather than at the intervention is not unique to India. The best-known instance is the Canadian Supreme Court’s 5-4 decision in Harvard College v. Canada (Commissioner of Patents), 2002 SCC 76, which allowed Harvard’s process claims and refused the claims to the mouse itself.
Section 3(c) already carries a High Court split on biotech eligibility
A split between two High Courts already exists on Section 3(c), the clause that sits beside Section 3(j) and catches much the same subject matter. It surfaced in a judgment the applicant cited here for Section 3(j), which in fact decided nothing about that clause. In Imclone LLC v. Assistant Controller of Patents and Designs, 2024 SCC OnLine Mad 8397, decided on 6 March 2024, the impugned order had recorded that the applicant’s arguments on that objection were found satisfactory and the objection had been waived. Justice Senthilkumar Ramamoorthy therefore ruled on Section 3(c), on a monoclonal antibody generated by replacing murine genetic material in mice with human material, injecting an engineered antigen, and fusing material extracted from the spleen with myeloma cells.
Three holdings came out of it. The qualifier “mere” in Section 3(c) is confined to the nearest reasonable referent, the discovery of a scientific principle, and does not extend to the discovery of any living thing or non-living substance occurring in nature. In that third limb, “occurring in nature” qualifies only the nearest reasonable referent, “non-living substance”. The exclusion therefore reaches only “the process of finding a hitherto undiscovered non-living substance by identifying and isolating it from nature”, and the Court directed the claimed invention to proceed to grant on the claims submitted during the hearings.
In reaching the first of those holdings the Court deviated from the Delhi High Court’s reading in Diamond Star Global Sdn. Bhd. v. Joint Controller of Patents and Designs, 2023 SCC OnLine Del 1879, so a divergence between two High Courts on the reach of “mere” is already on the books. The practical consequence is the one the Madras High Court stated itself: for a synthesised non-living substance such as a monoclonal antibody, the real challenge is establishing novelty and technical advancement rather than patent eligibility, and the applicant must still satisfy Section 2(1)(j). Whether the same is true of a cell is the question Section 3(j) holds open.
The Patent Office adopted the BTS test for Section 3(j) in its 2026 draft guidelines
The Patent Office has taken the applicant-friendly half of this argument into its examination practice and left the half that decides Harvard’s case alone. The Controller General published Draft Guidelines for Examination of Patent Applications in the Field of Biotechnology, 2026 for public comment in September 2026, alongside a companion draft for pharmaceuticals, with a fifteen-day window for suggestions. Section 16 of the draft carries Section 3(j).
What it adopts is the BTS Research test, quoted at length, including the passage on how much human intervention takes a process out of the “essentially biological” category and its reference to the European Patent Office’s Broccoli decisions. It adds three further authorities on the same point. Two are High Court decisions, Sakata Seed Corporation v. The Controller of Patents and Design, decided by the Madras High Court on 19 July 2024 in CMA(PT) No. 30/2023, and Speaking Roses International Inc. v. Controller-General of Patents. The third is the Intellectual Property Appellate Board’s 2013 decision in Monsanto Technology LLC v. The Controller of Patents and Designs.
On micro-organisms the draft states the position the Controller argued in Harvard. Although micro-organisms are excluded from the non-patentability list, a conjoined reading with Section 3(c) implies that only modified micro-organisms, which do not constitute the discovery of a living thing occurring in nature, are patentable subject matter. Claims relating to essentially biological processes of growing plants, germinating seeds, or the developmental stages of plants and animals are to be objected under Section 3(j).
What the draft does not do is give an examiner any illustration for a product claim to an engineered cell. Both worked examples under Section 3(j) concern plants: Biotech/IE-27, a genetically modified rice plant carrying a transgene for a drought-responsive transcription factor, which the draft treats as excluded regardless of the method of modification and points instead to the Protection of Plant Varieties and Farmers’ Rights Act; and Biotech/IE-28, a method of producing hybrid seeds by breeding a male fertile parent with a substantially male sterile one. The draft therefore supplies a test for process claims and no worked example for a product claim to an engineered human cell. Harvard’s amended claim 1 is exactly that.
Drafting a biotech claim against the Section 3(j) exclusion
Four things on this record bear directly on how a cell claim should be drafted and prosecuted in India. The composition wrapper is the first, because it attracted Section 3(e) and the Section 10 objections without buying any distance from Section 3(j), and the 2026 draft confirms that a composition framing is examined on what the active ingredient is. The functional label is the second: “non-native” needed a defined comparator on the face of the specification, and without one the claim stayed vulnerable under Section 10(4) whichever way eligibility went.
The third is the amendment itself, which has to be placed on the record in a form the Controller must address whenever it changes the nature of the claim. That is what Jitendra Kohli, Akebia Therapeutics and now the Harvard appeal all turned on. The fourth is the weight of foreign grants, and Harvard pointed to grants on corresponding applications in the United States, the United Kingdom, Australia, Germany, Japan, Korea and Russia. The Controller answered that Indian practice applies a stricter construction of Section 3(j), and BTS Research, the applicant’s own authority, puts foreign practice no higher than persuasive.
Nothing in the 2026 draft or in this appeal makes a refusal unlikely, and a Delhi High Court decision of February 2025 shows that the other direction is live. The Court there upheld the refusal of the Regents of the University of California’s application for a livestock vaccine built on a recombinant Salmonella, resting on Section 3(c) and on the disclosure and deposit requirements of Section 10, without reaching the micro-organism carve-out in Section 3(j) at all. For the Harvard file the only fixed point is the six-month direction in the remand.
Frequently asked questions
Did the Delhi High Court hold that stem cell derived cells are patentable in India?
The Delhi High Court decided no such thing. It set aside the refusal because the Controller had reasoned only on the original claims, and paragraph 20 of the judgment records that it had not considered the merits and that the application was to be decided without being influenced by any observation in it. The phrase “man-made construct”, which several reports attribute to the Court, appears in the appellant’s rejoinder at paragraph 8.10. Eligibility under Section 3(j) is for the Controller to decide on remand.
Does Section 3(j) of the Patents Act ban every patent on human cells?
Section 3(j) excludes plants and animals in whole or any part thereof, and the contest is over whether a cell produced in vitro is a part of an animal at all. The Controller reads the clause as covering parts however they are cultured, maintained or differentiated. The Calcutta High Court in BTS Research treated artificially produced tri-hybrid cells as falling outside the meaning of plants or animals, and the question has not been settled for cells derived from stem cells.
Can a micro-organism be patented in India when an animal cell cannot?
A micro-organism is carved out of Section 3(j), but only a modified one qualifies. The 2026 draft biotechnology guidelines state that a conjoined reading with Section 3(c) limits patentability to modified micro-organisms, because an organism merely isolated from nature is the discovery of a living thing occurring in nature. The Controller’s argument in the Harvard appeal was that this carve-out cannot be stretched to a genetically modified human cell, because a microbe is single-celled and an animal is a multicellular eukaryote.
What happens to a patent application after a High Court remands it to the Controller?
The refused order ceases to stand and the application returns to the Controller for a fresh decision, with the timetable and terms the Court fixes. In the Harvard appeal the Court required a detailed order on the amended claims within six months of communication of the judgment, after giving the applicant a hearing, and directed that the prior conclusions on the original claims not influence it. In Akebia Therapeutics the Madras High Court went further and directed that a different officer carry out the reconsideration.
Does a patent granted abroad help an applicant answer a Section 3(j) objection?
A grant abroad carries no more than persuasive weight on an Indian Section 3(j) objection. Harvard relied on grants for corresponding applications in seven jurisdictions, and the Controller answered that the reliance was misplaced because Indian jurisprudence applies a stricter construction of Section 3(j), particularly on human embryonic stem cells. Even the applicant’s own authority put it no higher, since BTS Research accepted that foreign practice is not binding and is persuasive mainly where it is consistent across several major jurisdictions.
References
Akebia Therapeutics INC v. The Controller of Patents and Designs, CMA(PT)/64/2024 (Madras High Court, 20 March 2025), as reproduced at paragraph 18 of the Harvard judgment
Association for Molecular Pathology v. Myriad Genetics Inc., 133 S.Ct. 2107 (2013); 2013 SCC OnLine US SC 47
BTS Research International Pty Ltd v. The Controller General of Patents, Designs and Trade Marks, IPDPTA 56 of 2023; 2025 SCC OnLine Cal 2943 (Calcutta High Court, 3 April 2025), as reproduced in section 16 of the 2026 draft biotechnology guidelines
Diamond Star Global Sdn. Bhd. v. Joint Controller of Patents and Designs, 2023 SCC OnLine Del 1879 (Delhi High Court)
Dimminaco A.G. v. Controller of Patents and Designs (Calcutta High Court, 2002)
Harvard College v. Canada (Commissioner of Patents), 2002 SCC 76 (Supreme Court of Canada)
Imclone LLC v. Assistant Controller of Patents and Designs, 2024 SCC OnLine Mad 8397 (Madras High Court, 6 March 2024)
Jitendra Kohli v. The Controller of Patents, 2022:DHC:1904 (Delhi High Court), as reproduced at paragraph 17 of the Harvard judgment
Monsanto Technology LLC v. Nuziveedu Seeds Ltd. (Supreme Court of India, 8 January 2019)
Novozymes v. Assistant Controller of Patents and Designs (Madras High Court, 20 September 2023), as reproduced in section 13 of the 2026 draft biotechnology guidelines
President and Fellows of Harvard College v. Controller General of Patents, Designs and Trade Marks, C.A.(COMM.IPD-PAT) 493/2022; 2026:DHC:2597 (Delhi High Court, 28 March 2026)
Regents of the University of California v. Controller of Patents, Patent Application No. 201717005699 (Delhi High Court, February 2025)
Sakata Seed Corporation v. The Controller of Patents and Design, CMA(PT) No. 30/2023 (Madras High Court, 19 July 2024)
The Patents Act, 1970 (Act 39 of 1970), Sections 2(1)(j), 3(c), 3(e), 3(j), 10(4), 10(5), 15 and 117
Draft Guidelines for Examination of Patent Applications in the Field of Biotechnology, 2026, Office of the Controller General of Patents, Designs and Trade Marks, published for comment in September 2026
The Protection of Plant Varieties and Farmers’ Rights Act, 2001 (Act 53 of 2001)
The Agreement on Trade-Related Aspects of Intellectual Property Rights, Article 27.3(b)
Bar and Bench, on the Delhi High Court’s direction to re-examine the Harvard application
SpicyIP, on the refusal in the Regents of the University of California application
Disclaimer
This article is for informational and educational purposes only and does not constitute legal advice. Whether a particular cell claim is hit by Section 3(j) depends on what the specification discloses and on how the Controller decides the objection in that file, and the question the Harvard appeal raised is still open before the Controller. An applicant facing an objection under Section 3 should take advice from a qualified patent practitioner on the claims actually pending.
Allow notifications
